Softbank’s Blurry Vision: Parts I-VI

Author’s note: This series of episodes on the Softbank Vision Fund was first published in podcast form between September and November of 2020, and written in the summer of 2020. Some of the facts on the ground (particularly w.r.t. Greensill) have changed since, so there may be some updates at some point in the future. The original episodes can be found here: https://physicalattraction.libsyn.com/size/5/?search=Softbank

Softbank’s Blurry Vision: How a $100bn Tech Fund Failed
Part I: Son’s Vision

Hi all.

This episode, in fine Physical Attraction tradition, we’re going to cover something a little different.

I want to go into the world of tech startups, venture capitalism, and to tell the story of Softbank’s Vision Fund, and how $100bn that was supposed to revolutionise the future of technology has actually gone towards boondoggles, bubbles, and bad ideas — and, potentially in some instances, something even darker than this.

This is not necessarily a particularly new story; I’m certainly not the first one to tell it and there were plenty of people calling this at the time or in advance, who deserve more credit. It does extend more widely beyond Softbank. But I’ve been reflecting on some of these stories, and some of these companies and trends that I’ve been following in the last few years. So partly I want to tell you about this stuff just because I’ve been thinking about it so much.

And I also want to tell you all this story here, because I think it’s important to realise just how irrationally exuberant the pre-COVID world was, how techno-hype has lead us astray, and how much money supposedly clever people invested in ridiculous ideas rather than genuinely preparing for the risks of the future, and maybe, maybe, maybe finally this time we can actually learn something from it.

So I’m going to tell you the story of Softbank’s blurry vision.

Let’s begin by explaining what Softbank is, and what its Vision Fund is.

Softbank is a huge international congolomerate and telecomms company — it’s the second-largest company in Japan. It was founded by Masoyoshi Son back in 1981, and he’s really taken on the mantle of one of these visionary founders of large technology companies that exploded to become massive in recent decades — think Steve Jobs, Bill Gates, etc.

And like all of these founders, there are some pretty incredible stories of the guy’s entrepreneurship that go right back to create and reinforce this sort of “myth of the founder” that increasingly crops up when we talk about tech companies. His grandparents were born in Korea, and his parents were squatters in Japan. He moved to the US at the age of 16 to get a high-school education. While in college, he invented an electronic translator that helped make him his first million dollars. He made a further million dollars by importing used video game machines from Japan — this, in the late 1970s when things like Space Invaders were just coming out and taking the world by storm.

The money from this, and from other investments, was what allowed him to first start up Softbank as a going concern. The rest, as they say, is history. Here’s an article that was written up in the New York Times back in 1995 about Son, where you can start to see the way that people typically report on the guy:

“WHEN he was trying to start his software distribution business in the early 1980’s, Masayoshi Son rented a huge booth at a Japanese consumer electronics trade show and offered software companies free space to display their products. The booth was mobbed, but the retailers that attended the show dealt directly with the software manufacturers, bypassing Mr. Son, the middleman.

“I probably made back one-twentieth of the cost of the booth,” Mr. Son told the Harvard Business Review a few years ago. “After that, many people were laughing at me. They said, That guy’s really dumb. He’s a nice guy, but dumb.”
There are probably a couple of things that Son is particularly famous for. One is his investment in a Chinese e-commerce company called Alibaba. He picked up a $20m stake in Alibaba back in 2000. This turned out to be a pretty good investment, as Alibaba ended up being the Chinese equivalent of Amazon, and when Alibaba eventually went public back in 2014, that initial $20m investment turned into $60 billion.

No one calls Mr. Son dumb now. They call him the Bill Gates of Japan. The 37-year-old entrepreneur has built his Tokyo-based company, the Softbank Corporation, into the leading distributor of software and a major publisher of computer magazines here, parlaying it all into a $1.5 billion personal fortune.”
A three-thousand fold return on investment is obviously not bad. It’s widely considered to be one of the greatest investments in history, and, earlier this year, Softbank’s total holdings in Alibaba were still worth around $100bn.

The other thing that Son is particularly famous for is the fact that not all of his investments, historically, have worked out. If you’re investing heavily in a lot of internet companies in the 1990s, your career is bound to run into a bit of a road-bump in the way of the dot-com crash of 2000, when the inflated expectations over the profitability of internet companies resulted in a massive glut of venture capital spending… and then a colossal market crash. Companies like Pets.com, which raised $85m in funding, ended up being worth $0 overnight. At its peak, it was spending hundreds of millions of dollars on warehouses and advertising with only tiny revenues to show for it. And there were plenty of other companies who had similarly fragile business models, ploughing through huge amounts of money and burning tremendous amounts of cash, often without even having substantial sales behind them, let alone making profit. That’s when you know you’re in a bubble.

Part of what Softbank did in the run-up to the crash was to develop a massive venture capital arm. In July 1999, a few months before the crash, they set up Softbank Capital Partners, which ploughed $1.2bn into various different online ventures. Webvan was one — raised nearly a half billion dollars from sources including Softbank, to disrupt groceries with an online service. Another online delivery service was Kozmo.com, which raised $300m. Softbank invested $100m in OptiMark Technologies, a company that was going to let people buy and sell shares online — they wanted to render the New York Stock Exchange obsolete. They invested $70m in a company called AllAdvantage — this actually paid users to install a toolbar which would gather data on them while they surfed the internet, with the idea being that they would then sell the data on to advertisers.

By November 2002, all of these companies had gone bankrupt and laid off all of their employees. And while some of Softbank’s other investments did manage to struggle their way through the crash, there were plenty that had to engage in mass layoffs to do so, and many whose lofty dreams of world domination never really came to fruition.

There were some big losers during the dotcom crash, but no-one lost more than Masoyoshi Son. Softbank’s share price crashed from an irrational peak of 10,000 yen in February 2000 to 154 yen in November 2002. For three days at the peak of those inflated expectations, Son was the richest person on the face of the planet, but it didn’t last. Son himself is said to have personally lost as much as $80bn in the dotcom crash, which would easily be the worst loss in history.

And although Softbank’s share price has recovered somewhat and Son himself is a multi-billionaire again today, with a net worth of around $25bn according to Forbes, he has never quite recovered those lofty heights from before the dotcom crash, when Softbank looked poised on the brink of world domination.

All of this background is to point out that making risky, high-stakes investments — sometimes gaining billions of dollars on an unlikely bet like Alibaba, sometimes losing untold sums in market crashes and when bubbles burst… this is nothing new for Masoyoshi Son. In fact, it has been the story of his whole life. And this is all there in the reputation of this guy — a visionary, a gambler, someone who has a “300 year business plan”, to whom losing the greatest amount of money in history is hardly enough to make you want to give up on grand schemes that would reshape the world.




You can get a sense of how Son sees himself — or, at least, how he would like to present Softbank and its “mission” — in a now infamous Powerpoint presentation he gave to the Softbank board in 2010, setting out this “30 year” and “300 year” vision for the company. I actually urge you to look this up, the whole Powerpoint presentation is online, and it’s well worth it.

The presentation is light on actual details about Softbank’s products and financials. Instead, it says: “Information Revolution — Happiness for Everyone!” and “Softbank comforts people in their sorrow.” One slide helpfully informs us the “the saddest thing in people’s lives is loneliness.” Another introduces Maslow’s hierarchy of needs. Son urges Softbank to “incorporate corporate DNA that will allow the company to survive for 300 years”, and “look into the far distance whenever you get lost.” There are, of course, lots of stylised graphs with exponentials on them, a projection that computing power will exceed the human brain in 2018, and Softbank is “working to realise” the “greatest paradigm shift of humankind”… all of this rhetoric will be very familiar to anyone who has listened to our episodes on the Singularity, the idea that artificial intelligence is going to take over from human intelligence in the near future.

There are lots of slides about love. Softbank hopes to allow people in Japan to live to be 200 years old, and communicate through telepathy… and of course, to co-exist with humanoid robots. Son points out that there are lots of threats facing humanity (although, oddly, doesn’t include climate change, but does include meteorites) but then pivots, saying “Yet people long for love, and are hurt by love.”

There are some actual, you know… tangible ideas in here. Son does say that the company should focus towards “strategic partnerships” rather than doing everything in-house, and argues that they should head towards buddy-like relationships with companies as well as capital relationships. He suggests a decentralized structure with lots of different arms to the business is more likely to be successful, which I suppose makes some sense. He wants Softbank to end up being involved with thousands of different companies over the next thirty years, which is where the Vision Fund comes in. And we get the five values of Softbank, which are:

Focus on information revolution and contribute to people’s happiness
Be ambitious and tenacious of justice
Commit to being an overwhelming Number 1
Think till our brains crush
No revolutions are down to the earth

So… it all explains a lot.

None of this is to mock Masoyoshi Son too much. It’s clearly important to have a vision beyond just what seems likely to make you money in the short term, or you wouldn’t have anything to work towards. And you have to assume, given his career to date, that the guy has a level of business sense and experience that mere mortals like us can only dream of.

But what I do see here is a hell of a lot of techno-woo. It seems like someone who has read up on a lot of futurist thought, a lot of wild speculation about how technology will evolve over the next century or the next few thousand years, who has read a lot of Singularitarian stuff where you just extrapolate exponential growth of technology in the last hundred years into the future, and who has bought in almost entirely to this very hand-wavey, very science-fiction, very idealised notion of how technology is developing into the near future… and who wants to rely on all of these techno-dreams about how technology is going to develop in the near-term… and turn them into a business plan. Oh, and he happens to have access to $100bn to spend on pretty much whatever he wants.

So it’s in this context that you have to understand the Softbank Vision Fund. When it was launched, back in 2016/7, it was a pretty colossal deal. It was going to be a $100bn fund that would invest in the next generation of technology startups and companies, looking for the next Alibaba type investment, and ploughing money into all of these futuristic technologies — artificial intelligence, Internet of Things devices with chips in them, and robotics companies.

To give you an idea of the scale here: when announced, that was the same size as all of the funds raised by US venture capital firms in the prior two-and-a-half years, and, investing their money over a five-year time horizon, it would represent a quarter of all VC investments in the US. So there’s this huge venture capital industry — pushing money into startups in the hope that some of them will end up being immensely profitable and paying back your investment many times over — and around a quarter of it was going to be just this one, titanic “Vision Fund.”

Son described it as follows:

“Technology has the potential to address the biggest challenges and risks facing humanity today. The businesses working to solve these problems will require patient long-term capital and visionary strategic investment partners with the resources to nurture their success,” Son said in a statement.

“SoftBank has long made bold investments in transformative technologies and supported disruptive entrepreneurs. The SoftBank Vision Fund is consistent with this strategy and will help build and grow businesses creating the foundational platforms of the next stage of the Information Revolution. “I make investments based on a vision,” Son said during an earnings conference Wednesday. “I believe the SoftBank Vision Fund is the first of its kind to be making nonstop, coordinated, vision-based investments of this scale. It’s a completely new animal.”

And Masoyoshi Son didn’t want to stop there. The plan was actually to raise another $100bn every five years — an eternal vision fund, if you like, that would just constantly pump money into the technology sector and into startups, looking for these bets that would actually pay off.

Where did the money come from for the initial vision fund? The Fund’s cash comes from multiple sources, with the Saudi Public Investment Fund the biggest contributor, having committed $45 billion to the project — so around half of the fund is oil money via the Saudi government, and a further $15bn comes from the UAE’s Mubadala Investment Company — which is also state-owned — with SoftBank itself contributing $25 billion. From the perspective of the Saudis and the UAE, the attractiveness of the Vision Fund is ostensibly that it allows them to diversify their economies away from being overly dependent on oil by getting some capital exposure to the global tech industry. Apple also pledged $1 billion to the Fund, and the remainder is made up from investments from companies including Qualcomm and Foxconn.

This funding has been described as “hastily assembled” — and one thing that is notable is that, unlike some other VC funds, the Vision Fund is not averse to going into quite a bit of debt to assemble this capital. This from Prospect magazine.

“The unique structure of the fund meant that the company quickly began feasting on debt to help boost returns. According to The Wall Street Journal, about 40 percent of the money promised to the Vision Fund by investors other than SoftBank takes the form of preferred stock, which promises a return of 7 percent a year regardless of the fund’s performance, just like debt. That structure is peculiar for a fund so focused on young, money-losing companies with no clear path to profitability, in the short or long term.

Under that structure, the fund’s stockholders get big returns on the way up. But for holders of the non-preferred stock, half of which belongs to SoftBank and its employees, the potential for losses on the way down is huge. A recent analysis by The Wall Street Journal showed that “the fund would need to generate around $12 billion in cash every year to produce a 20% return.”

The Vision Fund isn’t merely staked by money that functions like debt; it’s also got plenty of traditional debt as well. SoftBank’s $33 billion stake in the fund also relies on borrowed money, while the company itself has more than $160 billion in interest-bearing debt on its balance sheets. That figure, according to TechCrunch, is more than six times the amount the company earns on an operating basis, and just slightly less than the public debt held by the government of Pakistan.

That’s not the only peculiar trait of the Vision Fund. The company has extended about $5 billion in loans to employees to invest in the fund. Again in late September, the call went out to encourage employees to take out large personal loans to buy further into the fund, and prop it up with much-needed cash. Not only is the Vision Fund overextended, so are the people most intimately involved with it.

Some executives have been encouraged to borrow more than ten times their salary, according to the Financial Times, a move that is expected of its employees, as a test of loyalty to Masayoshi Son. Son himself has already pledged 38 percent of his shares in SoftBank as collateral for personal loans from 19 financial groups globally, including Mizuho, Credit Suisse, Julius Baer, and Switzerland’s J. Safra Sarasin.”

As I write this in 2020, the first round of the Vision Fund is mostly spent and Softbank are actually trying to get the second round of funding for the Vision Fund off the ground… but the COVID-19 pandemic, and some very particular impacts about what’s happened to the Vision Fund in the last few years, have cast the whole thing into some doubt.

I want to discuss in a little more depth some of the companies that Softbank Vision Fund has been investing in, and justify the title of this episode… but first, I want to tell you about how I first became aware of Softbank and the Vision Fund, because I think it, kind of unintentionally, illustrates a point.

The University of Oxford, where I studied (and still do) has very long summers. Typically you want to go and do something practical with all of that time. I took an internship on one of these summers to work for a tiny startup incubator which I won’t name, although I suppose it doesn’t really matter, because it’s all on the public record. And while it didn’t really lead anywhere directly in terms of my career, I learned some pretty valuable things from my time there. At that point in my life, three years ago, the Softbank Vision Fund was just about to launch, and I had just started the show you’re listening to right now.

The job that I had in this little startup incubator was to try and work up the details of a pet project that my boss had been dreaming about for many years. And it was a pretty wild one, too. Essentially, we were trying to work out if it would be possible to pitch — and achieve funding for — a programme that would look into humanoid robotics. The dream that my employer had was that it might be possible for a nimble start-up to develop a humanoid robot that would, essentially, one day, fulfil the role of the humanoid robots we all remember from science fiction: robot butlers, robots that can replace the work of humans, this sort of “last invention of mankind” that was going to render us all obsolete.

I have to be honest here. Quite quickly, it became pretty clear to me that the notion that we were going to actually pull this off was… extremely unlikely. But my job wasn’t as some kind of robotics expert; my job was to do an awful lot of research and compile reports that would let my boss know more and more about the state of the current robotics industry. At that point, I didn’t even know what venture capital was, what angel investors were, or the first thing about robotics.

In hindsight, it was a great job. I got to spend all day, every day, learning more about the tech industry, learning about venture capital, and educating myself about an entirely new aspect of the world that we live in. It probably won’t come as a shock to you that I love doing that kind of thing, and it’s informed a lot of my writing for Singularity Hub, where I ended up being a sort of climate-and-robotics expert for a number of years, alongside some of the topics we’ve discussed on the show about natural language processing and chatbots and so forth.

It was in this context that I first learned about Softbank. One of the most visible robotics companies out there at the time was “Softbank Robotics”, which produced the Pepper robot. [I later learned that this was actually an acquisition — another bet made by Softbank if you like — who had bought a French company, Aldebaran Robotics.]

[intercut some Pepper?]

You may have seen Pepper before — it’s this small, white, plasticky robot that wheels around the place and generally has a touchscreen in its belly. Go and look up some videos of Pepper in action if you want to see what I mean, and you haven’t seen the robot before.

The original aim of Aldebaran Robotics was to produce a robot that could help people who were sick or infirm. This has long been a big part of the practical motivation for pursuing robotics in Japan — alongside, of course, the cultural motivation that comes from Japan’s obsession with robots, which stretches all the way back to at least the Astroboy and Tetsujin-28-Go anime back in the 1960s. In practical terms, robots have often been invoked as a solution to the crisis of Japan’s aging population. By mid-century, over half of Japan’s population will be over the age of 65. There simply won’t be enough people of working age to look after them all, and to do all of the jobs that require you to be physically fit. The idea, then, is to develop robots that can look after elderly people.


There’s just one issue with Aldebaran/Softbank Robotics’ aim to produce a robot that can look after the elderly, though. They can’t make a robot that can do it.

The reality is that robotics, despite all the advances that we’ve seen, is still a long way away from being able to achieve the type of tasks that people have envisioned for it. Computer vision, and object recognition, has come on substantially in recent years. And there are some increasingly impressive precision robotics arms that are capable of manipulating objects. But neither of these are solved problems, by any means. And the abstract decision-making — to not only recognise objects, but also to recognise situations, and to know how to behave appropriately in that situation — I still feel that we are a long way away from that. And these things are nowhere near to actually being developed, or integrated in a single system that works or behaves in the way you’d expect it to.

One of the moments that demonstrated this to me was watching Robocup @ Home — this is effectively a robotics competition held every year where teams of university researchers and students try to get robots to complete basic tasks, like going shopping, cooking meals, tidying up, and so forth. I don’t want to disparage the efforts of the researchers at all — I know that I would have no hope of doing anything like what they’ve done! — but we take these things for granted, and we take our own innate intelligence as humans for granted. We assume it will be simple to programme robots to do similar things. In reality, though, when you watch these challenges — and see how stilted the robots are, how slow they are to move, how often they have to be prompted…. And most of all, how specialised and specific the tasks have to be, how clean the environments have to be, to avoid confusion for the robot… you can see that we are a really long way away from developing humanoid robots that can do anything useful, let alone emulate what humans can do.

This is why Pepper is essentially limited to being used as a gimmick to advertise Softbank. And as a gimmick, it works pretty well — you probably recognise Pepper if you read anything about technology or any “future-oriented” article about robotics or artificial intelligence, because they love using photos of Pepper almost as much as they love using misleading pictures of the Terminator. Yet there’s very little that the Pepper robots can actually do that couldn’t also be achieved by an iPad on wheels.

There’s a reason why, for example, the Boston Dynamics robots — like ATLAS, and Spot the Dog — who produce those viral videos where the robot does a backflip and so on… they have no practical uses anywhere: they simply aren’t ready to perform tasks, and it makes no sense to use them for anything practical when human labour is so cheap. In fact, this was the real moment when I realised that we were not going to achieve anything with the startup. Boston Dynamics was bought by Google as part of Project Replicant, which appears to have been an aim by Google to develop humanoid robots for some practical use. After five years, in 2014, they sold Boston Dynamics and Replicant was quietly disbanded. When I learned about that, I thought… there’s absolutely no way that we are going to be able to accomplish something here that Google gave up on, with all the programming resources and all the financial muscle that they can draw on to make such a project a success.

Of course, it probably won’t surprise you to learn that a few years after Boston Dynamics was sold by Google, it was bought by… Softbank! Softbank bought Boston Dynamics for around $100m back in 2017! Masoyoshi Son at the time said:

Today, there are many issues we still cannot solve by ourselves with human capabilities. Smart robotics are going to be a key driver of the next stage of the Information Revolution, and Marc and his team at Boston Dynamics are the clear technology leaders in advanced dynamic robots. I am thrilled to welcome them to the SoftBank family and look forward to supporting them as they continue to advance the field of robotics and explore applications that can help make life easier, safer and more fulfilling.

Now, it’s true that the Boston Dynamics robots are pretty amazing, but it’s also true that they are far more notable for producing a series of viral videos where the robot does a backflip or walks across uneven ground, etc. rather than any actual practical application. The practical application doesn’t exist yet, because the technology is not yet there. Despite what those videos illustrate, these robots are often remote-controlled — not autonomous in many applications — and they need to be carefully programmed to actually achieve most of the tasks that you see them performing in these videos. It’s the development of that software, and making that software general enough and capable enough that it can adapt to many different tasks and environments, that is the “hard part” of making humanoid robots useful.

For more on this, you can watch footage from DARPA’s Robotics Challenge online. People often argue that the first use case for humanoid robots is going to be in missions that are too dangerous for humans — “dull, dirty, and dangerous”. For example, in military applications, or in responding to emergencies or in disaster zones. They can then argue that you need the robots to be humanoid if they are operating in the built environment, which is tailor-made for humans. If you need to open a door, or climb a flight of stairs, your Dalek-like robot won’t work.

And yet if you actually watch the footage, you’re far more likely to see a robot falling over despite a whole team of operatives trying to control it live — or taking twenty minutes to painstakingly climb up a few steps. The technology is seriously impressive — and perhaps when they do another challenge in a decade or two it will be able to pass with flying colours — but right now, it’s nowhere near ready to perform anything really useful.

It was only a week ago — 28 years into the history of Boston Dynamics — that they actually put their first robot on sale; Spot the Dog, which is supposed to be for industrial applications, and which will set you back $79,000. It only went on sale a few weeks ago, in June 2020, and so it’s unclear whether anyone has actually found a useful application where this robot adds value, but I have a suspicion that they are going to struggle to find one, at least in the near-term.

The reason I’m dwelling on the humanoid robotics angle of Softbank is that I think it’s really indicative of where tech in general finds itself at the moment. We are in the midst of a period of massive technological hype — in fact, if anything, we may just be falling away from that peak of inflated expectations right now with COVID-19. It’s an era when you can produce a robot like SOPHIA, which is essentially just an animatronic puppet — and the hype is so real that thousands of people will apparently believe this robot is “sentient” or deserves human rights. It’s an era when you can stick the word “AI” onto a company, or a basic software solution, and double the value that you can get out of the company from investors. It’s an era where we are imagining self-driving cars, humanoid robots, artificial intelligence that runs or influences the world, and where we have a whole practically pseudo-religious movement, in transhumanism, of people who are expecting technological advances to arise in the next few decades that will make them immortal.

And in many ways, I think it’s an echo of the earlier dot-com bubble. Remember those technology companies from the late 1990s that we discussed. They actually weren’t terrible ideas. E-commerce is obviously huge now: it is the business model of the most valuable company on Earth. The idea that you could make money by gathering people’s data and selling that to advertisers — this is now the business model of the third and fifth most valuable companies on Earth, in the form of Google and Facebook, who provide us with a free service that costs an awful lot of money to run so that they can gather this data on us.

These aren’t terrible ideas. But just because you have a good idea, doesn’t mean that the time is right for it, or that you’re not necessarily running a terrible business that can’t survive and will burn unbelievable amounts of money in the process. These aren’t terrible ideas, but in the 1990s, when these other companies were trying to do it, the time was not right for these ideas.

And it feels like tech is at a similar stage now — vastly ahead of itself, in the midst of a hype bubble that isn’t backed up by the solid results of what the technology can actually achieve. And for me, humanoid robots are the absolute pinnacle of these inflated expectations at the moment. They are a perfect sci-fi technology — because everyone can imagine a humanoid robot, how it would act, and behave. We have examples of them throughout science fiction that people can draw on. People half-read endless headlines about AI and speculative TV shows or proclamations by figures like Elon Musk and Masoyoshi Son, or they see misleading demonstrations of these robots online, and they get sucked in to the notion that these things are way, way further advanced than they actually are. Far fewer people have a realistic idea of how close these technologies are to being useful — and whether they are even practical at all, more to the point.

It might be a few decades of slower, more sustained, and less hyped-up progress before any of these technologies become mainstream and ubiquitous like smartphones and social media are today.

This is very true when it comes to robotics, some of the more advanced applications of machine learning, especially when people call it AI, … to endless generations of mobile apps that are all hoping to be the next company that will be valued at a billion dollars for “disrupting” something that we all currently take for granted…

I think it’s also true when it comes to Internet of Things devices, blockchain, and self-driving cars. And these happen to be all of the areas that the Softbank Vision Fund is investing in. Softbank’s very visible investments in robotics companies have it positioned as the company that is most forward-thinking, most ahead of its time, and most visionary. But for all that’s written about them, for all the flashy images and demonstrations, and for all the over-inflated hype, there’s not a great deal of actual value there yet. In my view, lots of these investments are going to wind up being money down the drain, and the tragedy of it is that there are plenty of exciting and vital technologies that could be invested in which aren’t seeing even a penny of the Softbank billions. I think there’s a very real possibility that we are just past the peak of a large, over-inflated tech company bubble — and, when it pops, just as occurred in the 1990s, Softbank is going to end up having made quite a few bets that will look foolish in the long run.

And I think that COVID-19 is really going to be the turning point, the accelerant that is going to pop this bubble and, potentially, send us all into a “tech winter” where these inflated expectations come crashing down. There have already been a lot of discussions of this nature surrounding “why Silicon Valley has been unable to address COVID-19”, which is an interesting question I want to address in some future episodes.

And there’s an issue with bubbles. Some of the companies Softbank has invested in may have good ideas. But as appears to be the case with a number of companies backed by Softbank, others promised a lot only to turn out to be vaporware or running some kind of scam. Others have technology that is light-years away from being useful; others have horribly unsustainable business models that require burning huge amounts of cash to survive. And others are just the kind of terrible ideas that somehow manage to raise millions of dollars in an environment of irrational exuberance, when you’re at the peak of a hype cycle, before it all comes crashing down. And that’s the issue with bubbles: when something happens, and the music stops… say, something like a global pandemic and economic recession… they burst. And then we see behind the curtain.

Next episode, we’ll talk about the first of Softbank’s BIG bets — the bet that it made on a company that has successfully incinerated billions of dollars over the last few years.

Softbank’s Blurry Vision, II: The Big Bet on Uber and Self-Driving Cars

So last episode we talked about the Softbank Vision Fund and the founder, Masoyoshi Son, serial tech entrepreneur, gambler, visionary, luminary, possibly somewhat delusional, who has a penchant for ploughing billions of dollars into various ventures in futuristic technologies, not all of which succeed. The Vision Fund was established in 2017 and consisted of $100bn that they intended to invest over the next five years, before raising another round of $100bn that was due to start later in 2020/2021. It is, to put it mildly, a mind-bogglingly large venture capital fund trying to bet on “the future”.

So what has Softbank put its money into? There are a few big “bets” that get a lot of attention, and some others that are perhaps even more interesting… so let’s deal with them in order.

One of the first and most notable bets was into Uber, the ride-sharing company that you’re probably already aware of where gig-economy taxi drivers take you from place to place. When Softbank invested in Uber, in 2017, Uber was already a huge company — in fact, at that stage, there were already more Uber journeys than taxi journeys in New York. I remember reading about this investment and being a little… disappointed that the Vision Fund didn’t seem to have a great deal of vision to it?

You might expect an early-stage VC to find hundreds of smaller companies with new ideas or provisional technologies, perhaps spun out of universities, that might take 5–10 years to come to fruition. These would be small companies with a novel idea which may or may not ultimately be a success, and the VC might spend a few million dollars on each of these companies to allow them to progress their technology and develop it further. Softbank weren’t doing that, though. Instead, they ploughed — at first $7.7bn in December 2017 into Uber, which was already a huge start-up. They have since put another $2.3bn into Uber and its Advanced Technologies Group, meaning that $10bn or around 10% of the Vision Fund has gone into Uber alone.

So it’s worth looking at Uber in a little more detail. Despite being so ubiquitous across many major cities across the world — except, perhaps, in China where Didi Chuxing, a competing service, is closer to establishing the monopoly… which Softbank has also invested in to the tune of $500m — Uber has never actually made money. There’s a reason that the company needs these constant injections of venture capital, amounting to tens of billions of dollars over the last few years — it’s not profitable.

In fact, Uber is probably one of the least profitable companies of all time. In the first quarter of 2019, before anyone had heard of COVID-19 I might add, the company lost $5bn. At the rate they have been going lately, Uber would burn through Softbank’s entire capital investment in six months of losses. How on earth do you lose $5bn in three months, you might ask? Well, a lot of this specific headline figure was down to Uber giving away stocks as compensation to its employees after they “went public” (floated the company on the stock market.) But even when you take that into account, they are still losing billions a year, and lost around $8.5bn last year. How can this be?

The answer appears to be on Uber’s business model. They have this Silicon Valley attitude of “expand first, achieve profitability later”. In other words, their first aim is to corner the market for ride-sharing and private vehicle hire around the world. To do that, they have to be able to expand their operations to virtually every major city and town in the areas that they want to cover, regardless of whether or not their operations in that particular place are making money. And then, they need to undercut everyone who already exists in that space there to get people to switch. They can do this through ride subsidies, making rides cheaper than they otherwise would be, special offers to get people to start using the app, etc., and spending billions of dollars on marketing of course to attract drivers and customers. In fact, if you look at the second quarter of 2018, Uber spent almost half of the money that it made (after paying drivers, taxes, refunds, and expenses) on marketing alone, which helped it to lose nearly a billion dollars in that quarter.

If you take a city like New York, the taxicab industry has only really managed to be profitable there with some quite strict regulations on how it can operate — to the extent that, once upon a time, the special taxicab medallions that would allow you to operate as a proper taxi driver would change hands for millions of dollars.

The innovation of Uber isn’t necessarily even the app, which is nice software that connects riders to drivers, sure, but it’s actually the circumvention of these regulations on the kind of hours cab drivers can work, the way you have to treat your employees, the vetting and training that they need to go through, and the prices that they can charge for their service. This is what Uber’s innovation is — as the Trashfuture podcast so ably pointed out, lots of companies that poise themselves as tech companies are really just companies that circumvent regulations.

For example, Uber were hit in November 2019 with a $650m by the state of New Jersey in the US. Uber classes its drivers as “independent contractors” rather than employees, which allows it to avoid paying all manner of taxes. The state of New Jersey argues that they actually clearly are employees of Uber, and therefore unemployment insurance and disability insurance taxes are due. By posing as a tech company with a new business model, Uber may have avoided billions in taxes and benefits like this that ordinary companies would have had to pay to their employees in the past in order to operate in the same way.

But even to survive, Uber and similar companies find ways to undercut existing markets, quite often in this “lost-leader” way which requires a constant injection of billions of dollars of venture capital to be sustainable.

So people have criticised Uber, and Softbank for investing in it, because the plan — if there is a plan — simply seems to be to pump so much money into these companies that “disrupt” existing industries, until they can recreate the monopolies that previously existed for these services, and start to charge enough that they’ll make some kind of money doing so. Part of the issue is that while there isn’t a monopoly, they can’t charge whatever they want, and other companies — that are willing to operate in fewer markets, or burn money even faster — can compete. The cost to a consumer to switch from driving with one ride-sharing company to another is as small as downloading an app, and it’s not too hard for drivers to switch from one company to another that might offer some advantages in compensation, treating its workers slightly better, or paying more. The competition with other services like Lyft and Kapten — which also burn through huge amounts of VC money to keep going and aren’t profitable in their own right — is causing Uber to have to burn even more money just to keep up its own market share.












Looked at in isolation, Uber is a pretty amazing company. Its drivers often earn less than the minimum wage — an average of just $9 an hour in the US. On top of that, because they provide the gasoline and the car and the servicing and the MOT etc, and they bear the brunt of the car’s cost depreciating over time, many of the operating costs of running a taxi firm are shifted onto the drivers, a lot of whom are even worse off financially from the deal than they would be if they had taken on a different minimum wage job. And ultimately, by the way, if Uber had established the transportation monopoly they dreamt of, you can bet that its workers would be paid even less, as the company wants them to have no competing alternatives to go to.

Uber has made many attempts to avoid taxes and regulations in order to save money and has often succeeded in paying tiny amounts of tax; and yet it is spending so much on expanding its operations and marketing them that despite all of this, it still manages to burn through billions of dollars worth of venture capital money every year.

Arguably it’s just an elaborate mechanism to funnel VC money towards people getting slightly cheaper taxi rides, while exploiting drivers and circumventing the law. All the while, the only way it gestures to being profitable is arguing that transportation is a $12trn industry and one day Uber will account for all of it… which just seems a little bit insane to be honest. The ultimate idea behind that is that Uber will someday replace people ever owning their own vehicles, and all transportation will become “mobility as a service”, but how much money and how long are you going to have to burn it to bridge the gap between where the company is now, and that envisioned future? Of course, to do this, you have to compete with public transport, which for most journeys is intrinsically going to be cheaper to run at a profit than ride-sharing… and also capable of providing the capacity in a way that having a city clogged with cars is going to do. To say nothing of the environmental cost.

It’s not impossible that there is a version of doing what Uber does that can make money, particularly if it’s restricted to markets with quite high demand. But the reality is that a lot of the foundational assumptions of the business — that it could achieve a monopoly, and then set prices; that it would save a lot on overheads by not owning any vehicles; that it would someday achieve such a size and scale that it would save money due to economies of scale; that regulators wouldn’t be able to touch it and it could earn money that way; and that competitors could be driven out by establishing a monopoly on riders and drivers… all of these core assumptions look wrong, and so the business looks like a really bad bet. And this was before COVID-19 and the coming depression. COVID-19 has obviously massively reduced people’s willingness to take trips and use apps like Uber and will continue to do so into the future. How long are these massive VC firms going to be willing to pump billions into this strategy when it has no clear route to ever turning a profit now?

One of the ways in which Uber attracts a lot of venture capital investment is by focusing on its research and development; indeed, $1bn of Softbank’s investment went straight into the arm of Uber that is aiming to develop autonomous vehicles. The allure of autonomous vehicles for Uber is pretty obvious; you can cut out having to pay the pesky drivers altogether, and thus further undercut existing services, even though you now have to presumably transition to owning, maintaining and operating a fleet of cars yourself. [Because of how Uber works, the cost of owning, maintaining and operating the cars has fallen on the drivers, which is yet another way that they shift costs away from the company compared to a traditional taxi firm.] But it seems unlikely at this stage that autonomous vehicles are really going to be on the roads in force at any time in the next ten years or so. There are an awful lot of problems that have not yet been addressed.

For example, Uber itself had a famous fatal crash with one of its self-driving cars in Arizona in 2018. The incident report said, astonishingly, that “The system design did not include a consideration for jaywalking pedestrians” — and Uber’s self-driving cars have been involved in 37 crashes and other incidents in the last few years.
Looking at the signs, you can see that internally, privately, while a lot of self-driving car companies are still happy to sell you the hype that in the next few years we’ll all be driven around by autonomous taxis, they have actually pivoted quite substantially. Quoting from WIRED last year:

If you’re expecting self-driving cars to arrive soon. Prepare for disappointment. A decade of massive investment in robocar tech has spawned impressive progress, but the arrival of a truly driverless car — the car that can go anywhere anytime, without human help — remains delayed indefinitely. Despite Elon Musk’s self-assured claim that Teslas will have “full self-driving” capability by the end of 2020, the world is too diverse and unpredictable, the robots too expensive and temperamental, for cars to navigate all the things human drivers navigate now. Even John Krafcik, CEO of Waymo (the grown-up company that was Google’s self-driving car project), agrees, saying last year, “Autonomy always will have some constraints.”

That reality has pushed AV outfits to embrace “operational design domains,” engineer-ese for picking one’s battles. They’ll aim their tech at specific tasks it can handle, now or soon. The best way to understand the self-driving world is to ask not when it will arrive, but where. And how. And for whom.

This is a pretty familiar realisation… last episode I told you about how I was tasked with coming up with this business plan for a humanoid robot company. When it became clear to me how unlikely these robots were to be useful in the near-term, the best that I could do to try and salvage the business plan was to point out a whole bunch of “stepping-stone” industries — niches, essentially — that it might be profitable to invest in, or where the robotic technologies that exist today actually add value, which could eventually lead to businesses that could develop the technologies that would need to be incorporated into a more futuristic, humanoid robot. An example might be trying to repurpose a robotic hand to act as a pick-and-place robot in the Amazon warehouse. By focusing on this near-term, and hopefully profitable, application, you can actually work out some of the kinks of the technology in practical use-cases.

A classic example would be some of the incubation that solar panels got from their use on satellites. Solar panels were initially extremely expensive — but niche applications for where remote generation was required, or space-based applications like satellites, clearly were the kind of niche application where solar power had the edge. This allowed a market for solar panels that was just large enough for some serious research and development to occur that helped to make them cheaper in the long-run. Having realised that the dream of going from zero to a full-on robot butler was never going to happen with technology as it is today, I tried my hardest to come up with stepping-stone applications for each of the components of the robot, which would later lead to the whole.

[Incidentally, this is why I think that if any company will make serious advances in robotics, it’s going to be Amazon, who have a whole supply chain potentially where they can find these niche applications, trial them, and potentially perfect them fairly soon. But that’s another episode.]

And now we see that self-driving car companies are going down the same road. Self-driving shuttles might be restricted to particular routes, say, the train station to the airport, perhaps with a driver on hand in case anything goes wrong. Small spaces like university campuses, residential facilities and so on might be a good testing-ground for autonomous vehicles; they are easy to map out, they have predictable flows of traffic, and the speeds are generally slower. There are already long-haul trucking routes that are boring and predictable enough that they can be at least partially automated. Ditto some automated farming and mining machines where there is no traffic to worry about and you can essentially map out the entire route in advance if necessary — these applications have been working for the last five years or so.

I never want to deny technological progress or take away from the remarkable achievements that people have made in all branches of tech. I just think that it’s far easier to conceive of a lot of technologies working than it actually is to achieve them; people get seduced by the notion that you can do things a lot more easily, and that things are a lot closer to fruition, than they actually are. There are enough niche applications that self-driving tech is only going to continue to improve. But if it’s a decade, or two decades away — or even further away — from being able to replace an Uber driver in all but the most limited of circumstances, then it’s not really an argument for Uber’s business model. Because I doubt they can continue to keep burning cash at this rate for another 10, 20 years before self-driving cars come to their rescue and allow them to totally eliminate drivers from their business model and become more profitable that way.

And even if self-driving car technology does become mainstream, Uber would then have to spend a colossal amount of money to purchase a fleet of self-driving cars. Car manufacturers that are developing self-driving technology, who can manufacture the cars in-house, would seemingly have a much less costly route to replicate what Uber does with self-driving vehicles than Uber itself would. And Uber’s assets — the drivers, the passenger loyalty — have been demonstrated to shift pretty easily to competitors who can offer cheaper or more efficient services in local marketplaces. So it’s not clear what stops them from coming in and taking any profitability away from Uber, unless you have reason to believe Uber will be the only ones to get the technology working and it will be totally proprietary.

And I certainly think that, in Uber’s case, one of the main reasons for researching self-driving cars and making such a big deal about it is so that they can position themselves as a tech company, and continue to attract a lot of venture capital funding by discussing “disrupting” or taking over whole industries, when their actual core business model is to use that funding to undercut competitors, that “tech company” branding to circumvent laws and tax requirements, and eventually establish a monopoly when they can set prices. By calling themselves a tech company, and focusing on the digital side of the business, the hope is that they can persuade investors that they will genuinely be the next Amazon — companies that initially made quite heavy losses before eventually turning strong profits within a few years, due to the economies of scale that are much more realistic for a digital-based company than something like Uber.

So in a sense, you have to look at this as quite a predatory idea. The only route it has to become profitable is to undercut and eliminate the competition, establish a monopoly, and then jack up its prices and reduce its pay to the drivers, all the while circumventing regulation and taxes. Drivers have already seen their compensation, and initial bonus schemes, slashed in 2015/6. They have already been mislead by marketing material which claims they will earn much more from driving Uber than they actually will (when the costs of their vehicle ownership is taken into account).

If this was presented as the naked business model — a huge lost-leader attempt to establish a monopoly and dominate transportation — then perhaps fewer venture capitalists would invest. Instead, though, companies like Uber exploit the hype around technology — the belief that their app, or their access to “big data” or a proprietary algorithm — will make money. [If their proprietary tech was really so revolutionary, it’s not entirely clear how other companies have come in as direct competitors doing the same thing so quickly and in so many markets.] The belief that all it takes is for an “old industry” to be “disrupted” by a Silicon-Valley type tech company and endless profits will necessarily ensue. And the belief, perhaps most naively of all, that they can “disrupt” this industry with the rapid introduction of automated vehicles. It’s this techno-futurist, optimistic gloss on the business model that likely allowed them to sell themselves to a company like Softbank.

And it seems to have worked — because, after all, one wonders how much Softbank would have invested in them if they couldn’t have made that sort of pitch.

But, put quite simply, the technological innovation at the heart of Uber — the app that hopes to reduce wait times compared to cab firms — simply isn’t worth enough to justify the valuation of the company, nor the heavy subsidy on driving prices that it’s tried to use to get market dominance.
Because, while running Facebook does get cheaper per user as the company scales to have more and more users, the same is not true of Uber. 85% of its costs are for the drivers, the fuel, and insurance, and these costs don’t get cheaper when you scale to more users. And Uber’s attempts to expand and corner into many markets actually make their operating costs more expensive per mile than traditional cab firms that operated in individual cities.

Which leaves Softbank’s bet on them in a rather sticky position — as demonstrated by the fact that the Softbank Vision Fund has lost $5.2bn from its investment in Uber, as of May 2020, according to a Bloomberg report.

So Uber alone accounts for around $10bn of the Softbank Vision Fund. But when you add in Uber’s competitors — Grab, in Singapore, who got $3.5bn from the Vision Fund…

and on the tech side, Cruise, who got $4.6bn across two rounds of funding, and look to make self-driving cars… add in Manbang Group, who do the same thing but for trucks, who got nearly $2bn… add in Didi, who got $500m… and you see that in total, bets on companies that are Uber-like or rely on self-driving cars coming to fruition fairly shortly amount to around $21bn, a substantial fraction of the Vision Fund.

By the way — I should point out that none of these arguments about Uber or ride-sharing in general are new at all. This is the exact thing that people like Cory Doctorow and Hubert Horan have been saying for years — since way back in 2016/7, before Softbank even made their multi-billion dollar bet on Uber. In other words, it’s not like there was no warning for the Vision Fund in advance that this might not work out.

For example, this was Hubert Horan in May 2019.

“Few, if any of Uber’s narrative claims were objectively true. Hype about powerful, cutting edge technological innovations that could overwhelm incumbents in any market worldwide helped hide the fact that Uber was actually higher cost and less efficient than the operators it had driven out of business. Stories about customers freely choosing its superior products in competitive markets helped hide Uber’s use of massive subsidies to subvert market price signals and mislead investors about its growth economics. Misleading accounts about driver pay and working conditions helped hide the fact that most margin improvement was due to driving driver take-home pay down to minimum wage levels.

Uber was never going to dominate driverless cars and displace private car ownership, but these tales created false impressions about robust long-term growth. But all of these claims were uncritically repeated in the mainstream media, and over time they shaped the powerful general perception that Uber was “successful, efficient and highly valuable.”

The formula Uber used to build powerful, effective narratives was copied directly from what is widely used in partisan political battles. It combines significant resources (money and communication channels), the emotive, us-versus-them propaganda-style techniques demonstrated to be effective, and (Travis Kalanick’s contribution,) the willingness to deploy them in a ruthless, monomaniacal manner. The formula is especially effective when the interests that might disagree or challenge those claims were significantly less organized or funded.”

Horan’s original series trashing Uber’s economics came out in October 2016, and I urge you to give the 20-part essay a read if you need any more persuading that Uber was a bad bet.
So, again, this has been a long time coming. In it, he explains why the Amazon comparisons are misleading.

“It is useful to compare the public claims and perceptions about Uber’s growth with the case of Amazon, which like Uber, was seeking to drive a massive set of incumbent competitors out of business in order to achieve long-term industry dominance.

Amazon’s business model was focused on “disrupting” a book retailing industry that had high prices, high margins and high costs. By contrast, Uber cannot explain how it will realize billions in profit from an industry selling a commodity product with razor-thin margins that had already cut costs to the bone. Unlike Uber, Amazon proactively provided outsiders with compelling, verifiable evidence of the sources of its (potential) efficiency and scale advantages. These included the huge savings from eliminating “brick-and-mortar” retail locations, enormous scale economies in warehousing and distribution, sophisticated software that not only processed customer orders but dramatically simplified product search and identified customer-tailored buying suggestions, increased leverage with publishers and other suppliers, and huge scale economies that allowed it to expand geographically and into new markets at negligible marginal cost once its basic selling and warehousing/distribution infrastructure was in place.

The huge scale economies meant it could rapidly drive down unit costs as it grew, building strong loyalty through rock-bottom prices, and making it virtually impossible for existing (or new) entrants to ever match its efficiency levels. Amazon’s efficiency claims could be readily verified by objective outsiders who were expert in the relevant retailing, warehousing and ecommerce fields.

Amazon’s digital platform meant it could expand into other lower-margin businesses but did not invest heavily in these new businesses until it had secured a sustainable position in its core business. Unlike Uber, Amazon encouraged an active public discussion of its business model in order to build credibility and support in the financial community. While many observers were uncertain about Amazon’s long term profit potential, and questioned specific practices, there was universal agreement that its ability to rapidly capture share from industry incumbents was based on legitimate competitive advantages.”

You look at all this, you look at the fact that Uber claimed that it would be profitable in 2020 despite losing $9bn last year, and you look at the COVID-19 crisis which has massively hit its business model that was already losing billions… and it seems like the time is ripe for what a lot of people have been saying for a long time to happen, and for the bubble of these pseudo-tech companies to burst.

The final thing I want to mention in the tale of Uber as part of this series into Softbank’s Blurry Vision for its vision fund is the founder. Because Travis Kalanick, the founder, has been controversial for a long time — the sexual harrassment allegations, the depictions of a toxic workplace culture, his decisions to keep Uber as a private company for so long before it could go public and people would therefore get a better look at its financials and business model, and an ability to issue a verdict on its valuation…

Despite all of that, Kalanick sold his shares in Uber in December 2019, prior to what might be the final collapse of Uber, netting himself $3.1bn in the process for the invention of a company that has never turned investors a profit.

So obviously calling yourself a tech company can get you somewhere. Like any Ponzi scheme, it tends to work pretty well for the founder — and, indeed, any other investors who can liquidate their stock options and sell them to some investor in time, before leaving the company. Some people are doubtless going to make money from Uber’s over-inflated valuation. But it leaves an awful lot of other people holding the bag, and the Softbank Vision Fund is going to be one of the big losers.

I’m going to leave you with a quote from Byron Deeter, a partner in another VC firm, who is obviously revelling in the schadenfreude that comes with Softbank’s Vision Fund collapsing.

“The flood of SoftBank dollars caused some companies to pursue bad business practices with marginal returns, and they incinerated a lot of dollars in the process. I’m sure there will be valuable businesses that will come out of the fund, but this is an experiment that has largely run its course and is not going to be repeated.”

Next episode, we’re going to talk about another one of Softbank’s major investment bets — and another company that thought it could make a lot of money by posing as a tech company and “disrupting” an industry that really didn’t need disrupting… WeWork

Softbank’s Blurry Vision, Part III: The Big Bet on WeWork

Previously in this series, we’ve introduced Softbank’s Vision Fund, the $100bn VC fund poised as a massive bet into the future. We’ve introduced Masoyoshi Son, the founder of Softbank, and his delightful slideshows about how humans will soon be living side-by-side with robots and Softbank’s investment is going to accelerate the onset of the Singularity, and all that kind of thing.

And last episode, we’ve discussed how Softbank ploughed $20bn into various ride-sharing and autonomous vehicle companies, nearly half of that going to Uber, which poses as a tech company to get funding but has an arguably very shaky and predatory business model which relies on eating billions of dollars a year to continue operating.

This episode, we want to discuss another company that Softbank’s Vision Fund ploughed over $5bn into. If you add investments that were made by the parent company, Softbank Group, then Softbank has actually invested over $10bn just in this one company. It’s a company that posed as a tech company to justify an overinflated valuation, had a business model that required burning unbelievable quantities of cash over time… and if this all sounds familiar from last week, just wait until you hear what happened to the founder! I’m talking, of course, about WeWork.

Now you’ve probably heard of WeWork. If you know the full story of what happened to this company, which has been reported in a lot of different places already, then you probably won’t learn too much new from this particular telling of it. But it’s absolutely of a piece with some of the Softbank Vision Fund’s daft investments and a shining star in the tech bubble that was just about to burst before COVID-19, and is sure to be utterly destroyed now.

If you haven’t heard of WeWork, or only vaguely know this story, then please. Settle in for a wild ride.

What is WeWork? It depends on who you talk to.

Wikipedia, for example, says that WeWork is “an American commercial real estate company that provides shared workspaces for technology startups and services for other enterprises.” Essentially what WeWork do is rent out their own branded office space for tech companies to come and work in. They’ll buy up office space, put a great deal of branding all over it. They will provide the desks, maybe the computers, the Wi-Fi, the office furniture, fancy things like glass whiteboards you can write on, and pay for the upkeep of the space. And then, with all of these things taken care of, they will rent that office space out to people who want to use it, allowing them to concentrate on their work and providing that sort of short-term contract or ability to share an office with other companies that might usefully allow a smaller venture to get off the ground.

Okay, you might think — it certainly doesn’t seem impossible to make money in real estate. People need to rent offices. Perhaps that business model is going to be a lot more difficult in a post COVID-19 world where most office work is being done from home, and people don’t need WeWorks any more… but maybe… maybe before the pandemic, it was not a terrible idea?



And that’s when you find out how WeWork described itself in its S1 filing, the filing that companies have to produce when they want to go public and float on the stock exchange. The filing begins with an epigram “We dedicate this to the energy of We — greater than any one of us, but inside each of us.” The company is essentially a kind of intermediate landlord that comes into rental office spaces, brands them a bit, and manages the workspaces for tech startups.

It’s the business model of the tech bubble, that depends on a constant stream of people spinning the wheel of various industries, throwing a dart at it, and deciding to create a startup that will “disrupt” that particular industry with a mobile app or machine-learning algorithm or, more and more frequently, both. And, of course, it depends on all of those startups getting an endless stream of happy venture capital so that they can pay down their overheads towards renting the office space. In other words, WeWork takes in huge amounts of VC investment to operate itself, and relies on more VC investment to try and pay some of the rents for these properties.

But WeWork doesn’t want to be seen as a middle-man operator between landlords and other companies. Instead, they use the word “technology” over 110 times in the very vaguely-worded report about why you actually need WeWork and why you can’t just rent an office like a normal person.

“We provide our members with flexible access to beautiful spaces, a culture of inclusivity and the energy of an inspired community, all connected by our extensive technology infrastructure,” says the filing.

Why all this focus on “technology”? Again, there are a lot of similarities between Uber and WeWork. It’s not impossible to make money by getting people from A to B, in just the same way as it’s not impossible to make money by renting out office space. The issue is that they are pretending to be technology companies, and promising that they will someday attain some kind of monopoly on the market that they are entering into by being so “innovative” and “disruptive”. Their business model essentially consists of hoodwinking investors into thinking that they are going to be able to expand in a profitable way and make money from economies of scale like Amazon, when the actual fundamentals of the industry that they are trying to disrupt don’t lend themselves well to economies of scale. And, because they need to start by being lost-leaders to get the market share that they claim is necessary for the whole endeavour to work, these “visionary” companies essentially burn through billions of investors’ funds without ever making profit, before they eventually collapse.

It’s not necessarily that the service is bad. It’s just that these companies are coating themselves in this techno-optimistic, Silicon Valley, futuristic, exciting, and “disruptive” veneer. They spouting loads of nonsense about the company’s “vision” and “philosophy”, appearing innovative while not actually having any real innovations underneath. The reason they do this is that, in the boom times, this has been a really productive way of parting venture capitalists like the Softbank Vision Fund from their money.

As the tech websites Recode and The Verge pointed out back in August 2019:

“The We Company’s main competitor is IWG, a real estate company that is not pretending to be a tech company, as Recode points out. “IWG has had substantially more square footage and more customers, and has actually made a profit — yet its market cap (that is to say, the total value of its shares) is just 8 percent of what SoftBank’s latest funding round thinks WeWork is worth.” The We Company isn’t just a regular real estate company, then; it’s a real estate company that’s taken a lot of money from SoftBank and other firms by just saying “tech” a lot.”

If you look at both companies in 2018. IWG took $3.4bn in revenue and made $600m profit. WeWork took $1.8bn in revenue and lost $1.9bn in 2018. And yet WeWork was nominally valued at $47bn while IWG was nominally valued at just $3.7bn. Why on Earth is the more successful company here capable of being nominally valued at so much? Essentially because it called itself a tech company, and clever marketing, hype around technology and Silicon Valley and disrupting industries and other such jargon has caused investors to fling money at it. Chief amongst them, Softbank and Masoyoshi Son.

In other words, WeWork’s ability to benefit from this techno-hype — and the willingness of companies like Softbank to throw billions of dollars at companies that can successfully hype themselves up as part of this rosy, Silicon-Valley-ivied future — essentially caused Softbank to value them at ten times more than their competitor. All this despite the fact that their competitor actually makes money, has done for years, and on the surface of it is a much better-run and less risky business.

The story of how the Softbank Vision Fund came to invest $5bn in WeWork is really so extraordinary that I am just going to quote it in full from the article that Victoria Turk wrote about WeWork in Wired back in August 2018: “How WeWork became the most hyped startup in the world.”

“When Softbank first expressed interest in discussing business, Neumann insisted that Son should visit WeWork’s headquarters in person. “It wasn’t because we were trying to be cheeky,” he says. “It was because part of what we do is energy — and I can’t put energy on a piece of paper.” Initially they had planned for a two-hour visit and Neumann had prepared a presentation. That day, Son’s team first told him that Son was running late and that he would only have 90 minutes. They later corrected that to half an hour. “He comes in and he says, ‘I only have 12 minutes — go,’” says Neumann.

Unable to make it even as far as his office, Neumann whisked Son around the small R&D area and walked him around one floor of office space. Son then invited him to join him in his car so they could chat longer. During the trip, the two started sketching out a deal on an iPad.

Neumann says he would normally have waited a few days and tried to wrangle a better offer. But that morning he had attended a spiritual class, and his teacher had said that it is sometimes necessary in life to “do the opposite of our nature”. Neumann took the advice to heart. “I took the pen, I said, ‘Today I’m not a negotiator,’” he recalls. “I signed my name, his name, and that piece of paper ended up being the deal to a T.” The agreement with SoftBank Group and SoftBank Vision Fund, consisted of a $3 billion investment into WeWork and $1.4 billion into three new companies, WeWork China, WeWork Japan and WeWork Pacific. The deal cemented WeWork’s valuation of up to $20 billion.”

Now you might think that if you were going to invest $4.5bn worth of venture capital into a business — more money than pretty much anyone on the planet will see in a lifetime — it might be worth spending longer than 12 minutes talking to the CEO. If you’re busy, perhaps it’s worth delaying the meeting by a few weeks? It might be worth setting aside quite a considerable amount of time to consider the fundamentals of the business. It might, indeed, be worth doing quite a lot of due diligence to ensure that you weren’t just wasting a colossal amount propping up something that was fundamentally unprofitable. And I’m sure that there were other people involved in making the decision, and that it took more than just this meeting into account. And yet. And yet.

Is there anything behind the hype about WeWork? Well, it’s true to say that like Uber, it has grown rapidly into its new market. From 2016 to 2018, its revenue grew from $500m to $1.7bn. That might make you feel like the company is expanding, on the verge of becoming profitable. But Uber’s revenue has also grown year-on-year — it has also expanded into more markets and taken on more money — and all that’s happened is the rate at which it loses money has increased.

The issue is that almost all of this money is immediately spent on the bare minimum required to run this business. From a Bloomberg profile in August 2019:

“Last year in buildings that WeWork had up and running, the company recorded $1.5 billion in lease payments to landlords, plus costs for employees, utilities, real estate taxes, office cleaning, repairs and other expenses. That means WeWork’s revenue from operational office locations is scarcely higher than expenses for those locations — not including anything the company is paying for fixing up new locations that aren’t open yet, or costs for employees not working on operating office buildings…”

So, again, like in the case of Uber, we see that this isn’t really an industry with economies of scale. It’s not like a Facebook, Google, Netflix etc. where it barely costs them anything to serve another customer because all of the infrastructure, or content, or people to advertise to, or whatever you’re paying for, is already there. Every operation you run has this razor-thin profit margin because as you expand the business, your expenses expand in just the same way. There is nothing inherently cheaper about renting office space in ten cities instead of one.

This is reflected in the actual amount of physical assets that the average tech company needs to have. Facebook has $25bn of physical assets in terms of server farms, offices etc., but the company itself is valued at $525bn. It’s that kind of multiplier effect that allows you to grow rapidly without taking on debt. Perhaps when Facebook had only $1bn of physical assets it was valued at $20bn… which they can use to purchase more assets, and then grow.

WeWork could never do that because its actual value is not substantially more than the physical assets, in terms of leases, that it has purchased. The company would only ever be able to grow with substantially more capital to get more properties. And, of course, if there’s a market downturn and rents go down, the assets are worth less.

Tech companies can make money through “network effects”. Every additional user on Facebook or Twitter is more eyes on adverts; more content generated; more reasons for the next user to sign up. This simply isn’t the case with WeWork. There’s no reason that more people using WeWork would allow WeWork to provide a better service. Nor can it really leverage the data from its users to provide them with a better service, like Facebook or Google does by surveilling you and your behaviours to serve you ads, or, to be more accurate, to serve you to advertisers more effectively. [For more on how true that actually is, see our Patreon bonus episode on The Attention Merchants.]
What can WeWork do that’s equivalent? Learn how certain kinds of people like their office? That’s not the sort of effect that lets you print money in the way that other tech companies have.
WeWork has expanded very quickly to cities around the world, but it has done so by making a loss everywhere. It has managed to sustain itself by burning through billions of dollars to renovate these offices and purchase or lease them in advance, and undercut the existing competition. But by, somehow, positioning itself as a tech company and persuading certain people — mostly Masoyoshi Son, to be fair — that it will be the next Alibaba or Amazon and “disrupt” the office space industry, it’s managed to finance this undercutting expansion… until the music stops.

What is the “technology” that has allowed WeWork to describe itself as a tech company, rather than being honest and just admitting that it’s a real estate company?

Here’s an interview with Andy Palmer of WeWork, who was the VP of Product Development there, from early 2019 which aims to describe some of these innovations. He said:

“One of our most popular tools is the WeWork member app, which we recently redesigned from the ground up in order to better meet the needs of our members. Launching this app was one of my proudest moments at WeWork. Beyond helping members with daily tasks such as booking conference rooms, our app now offers features that allow them to connect with one another, whether they are in the same building or across the world. The app uses machine learning on the back end to match requests with appropriate members based on the skills and interests they’ve provided.”

So, um. The technological innovation that was supposed to make WeWork worth $45bn, and more than ten times the value of its closest market competitor which does the same thing but actually turns a profit… is an app that lets you book rooms. Seriously. The idea that it’s using “machine learning” to match you with other people with certain skills to aid collaboration is insane. You don’t need machine learning to do that. You just need a search function. The only reason that machine learning has been mentioned there is to make it seem like some magical, AI-enabled technology. I’m willing to bet that if ML is used at all in this app, it is totally redundant and could be achieved just as easily, if not much more cheaply, with a search function.

More to the point, LinkedIn can already do it, and it has a customer base of many millions more people than just those who work at WeWork. Why should the company that you rent your office from determine who you collaborate with? And even LinkedIn is worth substantially less than anything that WeWork has been valued at, and also appears to be a loss-making enterprise in itself.

At the peak of its inflated valuation, you can give WeWork people as much as you might want to explain why they are technology companies, but all you get is vague technobabble. Here is a spokesperson for WeWork quoted in the media when asked how they use technology: “We provide our members with space, community, and services through both physical and virtual offerings, all built on and powered by data, analytics and deeply integrated technology that help people unlock creativity and productivity. We use technology to help us deliver WeWork space more efficiently, build a global community, and improve the workplace experience for members around the world. For example, we built technologies to efficiently source, price and deliver locations around the world. These tools automate core tasks by moving manual management to digital systems, while allowing us to gather data at scale and apply machine learning to be more intelligent and expedient in our decision making.”

So, they, um… save money in deciding which offices to rent by getting “machine learning” to do it automatically? How much money does that really save when you have only a few hundred locations at most? It’s all just vague technobabble.

There were some further fun tidbits in a WIRED profile of the company from before its collapse in 2018.

“Sita shows me some examples of other technologies that WeWork is working on. There’s a smart phone-booth that knows when it’s vacant or engaged, and a door that unlocks when he holds his phone to it. He is particularly excited about a proprietary smart sit-stand desk. At the moment, the desk has a sensor underneath that can tell when someone is using it. The next step, he says, is for it to recognise who is using it and offer them personalised features. He waves his WeWork ID over a card reader in the desk’s surface and it automatically adjusts to suit his height. A fan plugged into a socket springs into life. “You’re going to give up a degree of anonymity and things will happen for you,” he says.

Another technology WeWork is experimenting with is facial recognition and sentiment analysis. Fano says facial recognition could be used as an extra level of building security or to “turbo-charge” its community managers — for example, giving the on-duty manager a heads-up that someone new has just entered the building.

When I approach a laptop running a sentiment analysis program in the demo area, it identifies my face and notes that I am female and wearing glasses. I make an awkward smile at the camera and get a high “happy” score. “We’re trying to uncover passive ways of understanding happiness,” says Sita.

He imagines cameras in a WeWork office tracking people’s moods via their facial expressions, perhaps combined with listening devices that analyse the emotion in their voice patterns.”

None of these technologies ever made it to mainstream use in WeWork companies, but… the best you can say about most of them is that they are going to allow managers to more effectively surveil their workers. As practical innovations, these are a long way away from the type of utopian vision for the “future of office life” and work that WeWork wants to sell itself as providing… instead, your desk will automatically spy on the length of your toilet breaks, and will offer you unspecified “personalised features” when you sit down at it.

It’s hard to see how any of this really rises beyond the level of novelty tinkering. And people did realise this at the time.

Quoting further from the Bloomberg profile:

“In short, everything about WeWork is utterly odd. It is a real estate company valued like a tech company. It is a young company with questionable economics that has committed to paying tens of billions of dollars in future years for office building leases.”

And this, again, is why investors should have looked at WeWork’s actual business model with a great deal of concern. WeWork is not a tech company. Their business model consists of renting office space (or occasionally buying it) from landlords for a long-term lease, and then hoping to make that money back by renovating the buildings and renting them out to their customers for lots of shorter-term leases, at a markup. This has worked fine for plenty of buy-to-let landlords, for example. But you can’t say it’s a recession-proof business model. If something were to happen that would suddenly mean that office space was a lot less valuable… like a global pandemic, for example… you are on the hook for all of these long-term leases with no actual short-term income from rents to pay for them.

But the actual business model didn’t matter. It certainly didn’t matter to the founder, Adam Neumann, the guy who was on the other end of the meeting with Masoyoshi Son. Quoting from a profile of WeWork (after its disastrous collapse) from the Atlantic:

“Adam Neumann’s overblown vision for The We Company is both the source of its success and the cause of its problems. Neumann has cited “energy and spirituality” as more relevant metrics for its potential on public markets than measures of its revenue and losses. The company was supposed to reinvent work itself, after all. Neumann’s leadership oversaw massive, rapid growth for the company — but it also seems to have been motivated largely by personal benefit rather than spiritual enlightenment.

The We Company’s Form S-1 outlined a labyrinthine ownership scheme that would have given Neumann more than half of the company’s voting power. Multi-class stock has caused problems for Facebook, Google, and other tech firms, but Neumann’s attitude was more brazen. He used his authority as CEO to pay himself $5.9 million for rights to use the new name, a change he had championed (he later returned the money under pressure of scrutiny before the IPO). He poured tequila shots for employees after announcing layoffs. He bought stakes in real estate that WeWork later leased. He declared WeWork “meat-free” by fiat in 2018, but failed to outline what that meant for its hundreds of offices and thousands of tenants. Rebekah Neumann, a WeWork co-founder and its former CEO’s wife, served as both brand chief and impact officer — a title that underscores the company’s proselytic ambitions — and also as CEO of The We Company’s private elementary school WeGrow, which is charging $42,000 in tuition for this school year (I swear I am not making this up).”

So you have to understand that a big part of this is really down to another thing that a lot of these pseudo-tech companies have in common, and it’s something else that has been extremely destructive as a stereotype, and that’s the myth of the founder. The fact that a set of founders — Steve Jobs, Bill Gates, Elon Musk, Mark Zuckerberg, and so on — have become so famous, so publically identified with their companies, and really hailed as visionaries who have reshaped the world… this has caused a lot of focus on companies with “charismatic” or “visionary” founders. In many ways, it’s caused far more focus on the charisma, unusual, or visionary nature of the founder than the actual, fundamental, technological advances that have been made and the way that the company intends to make money. Naturally this mythologising of how founders work has been written about very widely, but it is an extremely damaging stereotype.

You need look no further than Theranos — whose founder, Elizabeth Holmes, was the subject of a great deal of media attention both before and after Theranos collapsed. Their aim was to “disrupt” the diagnostics industry. Again, they preferred to brand themselves as a tech company rather than a medical diagnostics company. They had all the trappings of a Silicon Valley startup — the secrecy, the venture capital investment, the jargon, the branding, the appeals to lofty values like “family”, the desire to change the world. Holmes modelled herself on Steve Jobs to the extent that she dropped out of the same University and dressed in a similar fashion. But the whole thing was completely vaporware. Her diagnostic instruments didn’t work.

A lot of people were drawn in by the charisma of Elizabeth Holmes — and this was aided and abetted by the Silicon Valley myths, the idea that every industry on the planet is there for the taking, and that all it takes is one visionary (plus a few people who recognise the potential, and a whole tonne of slick marketing) to create a multi-billion dollar business. The technology, which was supposed to run many tests on a single drop of blood, was totally unreliable. When the fraud was discovered, the company fell in value from around $10bn to nothing. And most of the people involved ended up losing nearly everything. This story was told really well in the book “Bad Blood: Secrets and Lies in a Silicon Valley Startup”, which I recommend you go read. Enjoy the film when it comes out, probably later this year or next. Jennifer Lawrence is starring in it.

In hindsight stories like this should always be evidence for people that there is a wider bubble going on. I suppose the only reason Softbank’s Vision Fund didn’t invest in Theranos is that the company was already under suspicion as a huge fraud by the time the Vision Fund started spraying cash everywhere. But, amazingly, Softbank did invest in Theranos — after it had been exposed as a fraud — by acquiring all of its patents in December 2017 through its subsidiary, Fortress Investment Group, in exchange for $100m in cash. At last they had the sense to insist that it would be a loan, rather than an actual purchase, given that Theranos was being investigated for fraud by the SEC and had been exposed as having faulty technology months before.

The prominence of Adam Neumann and the founder myth that enabled him to work, which is really foundational to how Silicon Valley works, is a huge part of what has allowed this whole story to unfold. One profile of him in Vanity Fair by Gabriel Sherman — after the company’s collapse began — illustrated this well. Describing how Neumann was briefly worth $4.1bn, and spent lavishly on a $60m private jet and a $90m collection of homes, the article says

“In a way, the spending made sense, because Neumann himself was the product. He pitched himself to investors as a gatekeeper to the rising generation. A new way of working. A new way of living. Work was 24/7, coworkers were friends, office was home, work was life. For baby boomers who experienced office life as cubicles and bad coffee, his message was irresistible. “Every investor who walked through was sold,” a WeWork executive told me. They saw Neumann as a millennial prophet who did shots of Don Julio during meetings while preaching about the dawn of a new corporate culture, one in which the beer and kombucha flowed and MacBook-toting employees would love coming to work. After sitting with Neumann in his office, outfitted with a Peloton bike, infrared sauna, and cold water plunge, Steve Jobs biographer Walter Isaacson told Fast Company that Neumann reminded him of the Apple cofounder.

Through a combination of egomaniacal glamour and millennial mysticism, the Neumanns sold WeWork not merely as a real estate play. It wasn’t even a tech company (though he said it should be valued as such). It was a movement, complete with its own catechisms (“What is your superpower?” was one). Adam said WeWork existed to “elevate the world’s consciousness.” The company would allow people to “make a life and not just a living.” It was even capable of solving the world’s thorniest problems. Last summer, some WeWork executives were shocked to discover Neumann was working on Jared Kushner’s Mideast peace effort.

In April 2012, Neumann secured his first venture capital funding from Benchmark Capital, the San Francisco-based firm famous for its bets on eBay, Twitter, and Uber. Benchmark cofounder Bruce Dunlevie joined WeWork’s board after touring a WeWork with Neumann. “Bruce walked in and he said, ‘You’re not selling coworking, you’re selling an energy I’ve never felt,’ ” recalled a former WeWork executive who attended the meeting.

JPMorgan CEO Jamie Dimon was also an evangelist. In 2018, JPMorgan led a $700 million bond offering for WeWork. The bank extended Neumann nearly $100 million in loans and was among a group of banks that provided Neumann with a $500 million personal credit line.

Neumann’s most fervent believer, though, was Masayoshi “Masa” Son, the 62-year-old CEO of Japan’s SoftBank Group. In recent years, Masa had transformed SoftBank from a telecom conglomerate into the world’s most aggressive start-up investor, doing more than anyone to inflate the unicorn bubble. Backed by $60 billion from Saudi Arabia and Abu Dhabi, SoftBank pumped billions into fast-growing but money-losing companies like Uber, DoorDash, and Slack. Masa would invest $10 billion in WeWork. “Adam later said, ‘I’m crazy but Masa is crazier,’” a former WeWork executive recalled.

Achieving this blistering growth resulted in barely controlled chaos inside the company. “The place made the Trump White House look like a well-oiled machine,” a former senior executive said. The company voraciously leased office space with seemingly little strategy, except to keep adding locations. To entice new customers, WeWork offered members free rent and bought out their existing leases. “We are in a consumption phase, like nothing that has ever been seen,” Neumann declared at an industry conference in 2015.

The promise of IPO riches kept many employees from simply quitting. “The numbers they threw out at all-hands meetings was that this is going to be a multibillion-dollar company,” a former employee said. The money was only part of it, though. Neumann inculcated in his postcollegiate staff a belief they were members of a vanguard changing the world.

Neumann’s charisma was intoxicating to be around. “If you had to go to war, you wanted him to be your general,” a former executive said. “His sense of himself is beyond human,” recalled another. Neumann paraded through the office barefoot with celebrities like Drake and Ashton Kutcher and had an unnerving ability to maintain eye contact during conversation, lending him the aura of a guru. “When you’re in a room with Adam, he can almost convince you of anything,” a former employee said. Neumann used mass gatherings to spread his gospel. “I think the thing that all of us know is that if you want to succeed in this world you have to build something that has intention,” he said on stage at Summer Camp in 2013, his hair pulled into a ponytail. “Every one of us is here because it has meaning, because we want to do something that actually makes the world a better place. And we want to make money doing it!” The crowd of thousands exploded in cheers. “So many of the people were young and had never worked in a real company. They bought all of it,” a former senior executive said. “I realized after I got there it was a cult.”

Neumann’s gravitational pull was drawing in the world’s biggest investors too. In the spring of 2016, Neumann met Masayoshi Son at a dinner. The following year, Neumann invited Masa to WeWork headquarters. Masa informed Neumann he had precisely 12 minutes to hear a presentation. Afterward, Neumann followed Masa outside to his car, hoping to continue the pitch. Masa told Neumann he didn’t think his business plan would work. Neumann’s problem was that he needed to think bigger. WeWork shouldn’t just be leasing offices to small businesses — it should be leasing office space to all business. Masa scribbled on an iPad the number SoftBank was prepared to invest: $4.4 billion. “Adam dazzled Masa,” an investment banker close to both men told me.

Pumped up by SoftBank’s billions, Neumann’s messianism became more like megalomania. “Adam’s fantasy land became a reality,” a former WeWork executive said. Neumann sat down with world leaders, discussing the Syrian refugee crisis with Canadian prime minister Justin Trudeau and urban planning with London mayor Sadiq Khan. “When Adam got in front of world leaders, it was like he started thinking he was one,” a former executive said.

In conversations with people inside and outside the company, Neumann’s pronouncements became wilder. He told one investor that he’d convinced Rahm Emanuel to run for president in 2020 on the “WeWork Agenda.” Neumann told colleagues that he was saving the women of Saudi Arabia by working with Crown Prince Mohammed bin Salman to offer women coding classes, according to a source. In another meeting, Neumann said three people were going to save the world: bin Salman, Jared Kushner, and Neumann. Shortly after the news broke in October 2018 that Saudi agents tortured dissident and Washington Post columnist Jamal Khashoggi and carved his body with a bone saw, likely on order from the crown prince himself, Neumann told George W. Bush’s former national security adviser Stephen Hadley that everything could be worked out if bin Salman had the right mentor. Confused, Hadley asked who that person might be, according to a source familiar with the meeting. Neumann paused for a moment and said: “Me.”

Reality came crashing down last August when WeWork filed its S-1 prospectus to go public. Investors were shocked by WeWork’s spiraling losses and that the company had spent millions on Neumann vanity projects such as a wave-pool company and a start-up that sold turmeric coffee creamer. Most damaging, however, were disclosures that Neumann made $6 million by selling the “We” trademark back to the company and held ownership stakes in buildings WeWork leased from, essentially paying himself. (After criticism, Neumann returned the trademark money.) He gave his wife a large role in choosing his successor.”

So this is really the amazing thing about WeWork is that it managed to completely collapse before COVID-19. The company initially wanted to float and go public at its private valuation of $47bn, which is what Softbank valued the company at according to its investments. But investors in general were extremely wary about Neumann’s own personal dealings with the company, and its general lack of any kind of sustainable business model. Once the raw details of how much money the company was losing — as well as some more rumours about Neumann’s own self-dealing in the company became clear, people lost confidence. It didn’t help that the Wall Street Journal was reporting that Neumann was talking in private about becoming the world’s first trillionaire, the first person to live forever, becoming President of Israel, and also declaring that WeWork could be the company to end world hunger. They couldn’t find people willing to buy the company’s stock at even a $10bn valuation. And so the IPO — initial public offering — was shelved, which was the start of WeWork’s massive collapse.

After the IPO was shelved, Softbank essentially took the reins and forced out Adam Neumann. On September 23rd, the Washington Post reported that Softbank had “lost confidence” in his leadership, and on the 24th Neumann announced that he had stepped down as CEO.

In October 2019, in the wake of the disaster, Softbank revealed that they would provide another $5bn to WeWork in order to take 80% ownership of the company. This from Bloomberg:

“SoftBank Chairman and CEO Masayoshi Son said in a statement, “SoftBank is a firm believer that the world is undergoing a massive transformation in the way people work. WeWork is at the forefront of this revolution.”

The Japanese billionaire said WeWork’s “growth challenges” are “not unusual for the world’s leading technology disruptors.”

“Since the vision remains unchanged, SoftBank has decided to double down on the company by providing a significant capital infusion and operational support. We remain committed to WeWork, its employees, its member customers and landlords,” Son said in a statement.”

In November 2019, the company laid off around a third of its workforce in an effort to try and scale back and become profitable. And Softbank revealed that they had lost $9.2bn in investing in WeWork — almost 90% of their total investments over the last few years, since that twelve-minute office meeting and iPad-thrashed-out deal.

In other words, Softbank bet big on a real-estate company that posed as a tech company. It was able to say all the right words, push forward all the right branding, leverage the myth of “disruption” and the maverick founder who changes the world. It just didn’t have a viable business and was, if anything, valued at least ten times higher than companies in the similar industry that actually made a profit. Softbank bet big, and Softbank lost.

One person who famously didn’t lose out in all of this, though, was Adam Neumann. He got at least $1.7bn as an exit package for being CEO, with his $500m in debt to JP Morgan written down, $1bn for his shares, and hundreds of millions more in “consultancy fees”. While his former employees are out on the streets, and Softbank foots the bill, Neumann will never want for anything again.



Or, at least, that’s how it all looked pre-COVID. But as COVID-19 further destroyed any semblance of a business case that WeWork might have had for itself, things went from disastrous to worse, and the latest news is that Softbank is now trying to pull out of the deal to buy Neumann’s shares.

SoftBank said it had “no choice” but to scrap the rescue deal because WeWork had failed to meet several conditions. It also cited concerns about “multiple, new, and significant pending criminal and civil investigations”.

The Japanese conglomerate said it remained “fully committed to the success of WeWork” but “several of those conditions were not met, leaving SoftBank no choice but to terminate the tender offer”.

Neumann is now trying to sue Softbank, since them scrapping this rescue deal now means that he may never get his $1bn in shares (which may well be worth less than zero at this point) sold to Softbank. We’ll see how that goes. I wouldn’t hold my breath.

Uber. WeWork. Ross Barkan, back in the Guardian in Nov 2019, made a valuable case that links these companies beyond the fact that Softbank invested in them.

“WeWork is a junior cousin of Uber, another tech giant that has barely made a cent in profit. Uber is far more nefarious, driving down the wages of drivers and clogging city streets with unsustainable levels of traffic. It’s an unregulated cab company helmed by vulture capitalists. WeWork is comparatively benign: a lousy real estate company, conceived by Neuman, that tricked a lot of people into thinking it was something more.”

There is one last thing I want to say about WeWork before we wash our hands of it. Remember that I told you about IWG, the competitor to WeWork that was already established and profitable providing pretty much exactly the same service?

The fun thing about that is that IWG started off as an unprofitable company that was burning money, too. The rental agreements that they entered into, at the height of the dot-com boom in the 2000s, also ended up being wildly unprofitable. After that bubble burst, IWG was forced to file for bankruptcy, before gradually restructuring itself to become a smaller, less ambitious, and more profitable organisation. And lots of early venture capital investors in IWG also lost money.

So… not only is the current bubble that’s bursting in many areas of tech very reminiscent of the dotcom bubble… not only did Masoyoshi Son personally get massively wrapped up in the dotcom bubble and lost billions due to overinflated expectations about technology in the dotcom bubble… but two of the major losers from those bubbles had exactly the same business model! IWG is quite literally the ghost that haunted WeWork.

And technological hype, the mythology that has arisen from the relatively few titanic successes of Amazon, Facebook, Google, Netflix, has allowed people to defy the laws of gravity and continue to lose billions and billions in a notional “growth phase” towards a profit that never arrives. Barkan says that “Capitalism has always been a deeply flawed way of arranging a society, but big business, until around the 21st century, was at least bound by a few basic laws of monetary gravity.” Companies like Ford or US Steel could not lose money for years and years on end and still continue to exist with VC pumping money into it.

So far, then, we’ve talked about how Softbank’s two largest bets have essentially been into businesses that have posed as tech companies, but whose fundamental business models were flawed — and how both of these bets have lost the Vision Fund billions of dollars.

Can it get worse from here? I think the answer may very well be “yes, yes it can.” In the next episode, we’ll talk about a whole plethora of companies that Softbank has backed, and why they are also part of this vast techno-hype bubble.

Softbank’s Blurry Vision, Part IV: I Identify as a Tech Company

So in the last few episodes, we talked about how Uber and WeWork have both massively inflated the valuation of their core business by claiming to be “tech disruptors” and technology companies. And we talked about how this allowed them to extract a great deal of funding from Softbank’s Vision Fund, which seems all too happy to invest in anything that has branded itself as a tech company.

This is certainly the model for two of Softbank’s biggest investments. But actually, there’s similar nonsense in a lot of their smaller bets as well.

When it comes to these smaller bets, you start to get a feel for the type of companies that the Vision Fund typically likes to invest in. There’s a class of companies that might not actually be terrible investments, which essentially consist of Masoyoshi Son attempting to repeat his major success. By investing in Alibaba as a sort of China-based competitor to Amazon, which was founded a couple of years after Amazon first went public, there’s essentially an acknowledgement that Amazon’s business model has been wildly successful and that imitating it in a different jurisdiction may prove to be just as successful. This is I think behind some of the Softbank “bets” on companies like Rappi, which wants to do rapid delivery of consumer products in South America, or Flipkart, that is aiming to be the Amazon of India, essentially. About $2.5bn from Softbank has gone into Flipkart, and another $1bn into Rappi.

Now you can certainly criticise each of these businesses — they have had litanies of complaints about poor customer service, fraudsters exploiting the platform, and so on. But actually, the basic idea — to replicate what Son did with Alibaba — has probably led to some of the better Vision Fund investments. Flipkart, for example, was bought up by Walmart for $17bn — I don’t know whether Softbank made any money on that deal, but it certainly hasn’t seen the spectacular collapse of some of the other companies in the Vision Fund. If all Softbank was doing was investing in lots of companies like that, which aimed to bring proven business models to new, growing (geographical) markets like India or Latin America, then… okay, you can’t exactly call it a glorious vision of the far-flung future, but it might at least be a reasonable thing for venture capital to do, because if one of them did manage to become the next Alibaba, the investments would ultimately be justified in the same way.

Unfortunately, though, some of the other investments made by Softbank either fall into the trap of overvaluing a company that does something basic, like WeWork or Uber really do, because it has the trapping and brandings of a tech company.

Take Wag, for example. Wag is a company that essentially wanted to do for dog-walking what Uber did for ride-sharing. Again, dog-walkers are hired as independent contractors, and what Wag does is make an app that can connect you to someone nearby who will walk your dog for you.

Naturally, the issues that the company has had are very similar to that of Uber. In the push to go to more and more locations, they’ve burned through a lot of money without becoming profitable, undercutting existing services to try and get that monopoly on the dog-walking industry. Because the core “technology” of the company is so easy to imitate, they are in a race to the bottom with competitors, like the “Rover” app, which can do exactly the same thing. Consumers and dog-walkers alike can both switch between the apps and leverage whichever one is burning the most cash to attract their business. Again, much like Uber and WeWork, the margin on this dog-walking business is already razor-thin, and it’s not clear that you can both undercut the competition in terms of prices, pay a sufficient salary to attract dog-walkers, and generate enough spare cash to pay for a middleman company that was (after Softbank injected $300m into it) valued at nearly a billion dollars, trying to expand endlessly into other markets and also at a loss-lead.

And, of course, because we’re talking about people’s beloved pets, anything bad that happens (dogs running away, or being abused, etc.) is extremely negative PR for the company. It’s a little hard to try and counteract that with lots of marketing that relies on influencers on Instagram and Twitter, celebrities etc., to advertise how they are using Wag — because the bad-news stories surrounding the company will go viral much more easily.

Wag actually only asked for $75m in investment. But Softbank gave them $300m because apparently they are pathologically addicted to inflating a tech startup bubble for ideas that are likely never to actually turn a profit. And, at this stage, it’s not going to surprise you to hear that Wag has since laid off most of its employees, and Softbank has sold their stake in the company back to Wag at a pretty massive loss. We don’t know exactly how big that loss was, but it seems certain that Softbank essentially gave Wag that $75m, and more, to burn.

And ultimately, it really shouldn’t be this easy for you to just open the dictionary at a random page, and say, okay — I’m going to develop an app that’s going to be the Uber of on-demand hairdressing, or massages, or hiring musicians, or whatever — and then go to Softbank and convince Masoyoshi Son to pump your company with hundreds of millions of dollars. Yet, for the last 5 years, pre COVID-19, this has not only been a perfectly workable scam but probably one of the most reliable ways to make money. And the Viner brothers, who founded the Wag app, moved on from the company in 2018 to found a company called Wheels, which did the exact same thing but with electric bikes, earned $100m in venture capital investing, and directly competes with a service offered by Uber to do something similar. I am sure you will be shocked to learn that the last story I can find about Wheels, from February 2020, was that they were laying off staff in an attempt to become profitable.

One aspect of Wag’s story in particular that interests me is how this money actually goes to further inflate the hold of the existing tech giants on money and power. A lot of the software for wag is run on Amazon Web Services. A lot of their advertising and marketing money goes straight into the advertising black boxes of Google and Facebook… and, they also rent office space, although I can’t actually confirm whether or not that was from WeWork. We have discussed already how a lot of the advertising on Google and Facebook may be flawed and not providing a great deal of value [in our Patreon bonus episode on The Attention Merchants.] So if online advertising is also an overinflated bubble at the moment, then it is a bubble that the tech startup bubble inflated by Softbank has also, in its turn, helped to inflate.




Another startup that saw an injection of cash from Softbank was Zume Pizza. This is a particularly wild tale of excess. Remember when we first introduced Masoyoshi Son, and his PowerPoint presentation to the Softbank board? The tales of humans co-existing with robots that would serve us and bring us great happiness, perhaps eventually replacing us entirely? And the Vision Fund is supposed to achieve this, well, this vision of robot-human coexistence.

What you actually get, though, because of how far off this vision is from reality and how much fundamental research needs to be done before you can actually get there, is stuff like Zume Pizza. This was a startup that was going to use a robot, automated van which would automatically cook your pizza en route to your house. Again, the whole thing is quite reminiscent of Boston Dynamics in that it makes for a great, viral video but it’s not entirely clear how you are going to be able to use this to compete with… the hundreds of other businesses that deliver pizzas and are already established. What makes it superior, and what justifies the massive outlay of cash? Isn’t it just a gimmicky boondoggle?

Obviously not. Bloomberg had the inside story on how Zume Pizza managed to get $375m from Softbank’s Vision Fund. Naturally, a huge part of it comes down to these unbelievably inflated Silicon Valley tropes — a charismatic founder on a mission to “change the world.” You can imagine the pitch meeting, too, because at this stage, it’s actually very formulaic. You identify some industry that already exists. You point out that it’s worth x billion dollars a year. You project, without much explanation or detailed strategy, how with enough investment, your app could account for y% of those billion dollars in a few years time. And thus, of course, the startup will be a huge, disruptive success.

“If a founder seems brash enough, charismatic enough, reminiscent enough of a younger Son, some of SoftBank’s rigorous ­business-model tests appear to melt away.”

In the fall of 2018, SoftBank delivered its Vision Fund cash. After one conversation with Son, Garden choked up relaying the details to a confidant, saying, “Masa says I’m going to change the world,” this person recalls.”

It remains unclear to any rational person how, even if the business somehow worked and made money, having a pizza that’s cooked en route to your house by a robot is going to change the world. Imagine the number of research projects you could fund with $375m — think of the problems that need solving in the world, from diseases, medical therapeutics and diagnostics, vaccines, climate change, responding to natural disasters, and so forth. Theranos was a scam, but at least their dream was actually worth achieving. How does anyone really benefit from automated pizza delivery? Even if you want to have your pizza cooked by robots, since you are going to have to assemble the ingredients from some central location at some point anyway, how is it any cheaper to have a hundred vans with robotic ovens in them than a single, central robot oven that cooks all of the pizzas?



Not that Zume could even do automated pizza delivery. In 2018, IEEE Spectrum reviewed the company with “pizzas only a robot could love”, criticising the taste, lengthy wait time, and the price.

This is from a Business Insider article about Zume:

Business Insider toured Zume’s production area in Mountain View in August 2018, prior to SoftBank’s investment in the company. At the time, the robots were essentially mechanical arms that “press mounds of dough, squirt and spread sauce, and lift pizzas in and out of the oven, in a fraction of the time it would take human workers to do the same.”

It’s not clear how many of those robots were in existence and how that figure has changed in the years since, but a Zume spokesperson told Business Insider in 2018 that its Mountain View production facility was capable of churning out about 370 pizzas an hour.

Behind the orchestrated photo-ops, the real picture was not as pretty.

A former employee told Business Insider that the robots on which Zume, and Garden by extension, had built the entire business were inherently flawed and could not clear food safety inspections because they produced metal shavings, which could end up in finished pizzas. Business Insider confirmed the shavings were present in the food preparation area in a series of photos provided by the former employee.

“They had hired many people who had no idea what kind of equipment was supposed to be used in food, and many engineers were not food hardware engineers,” the former employee said.

With the robotics division struggling to live up to expectations, Garden came up with a new solution: delivery trucks that could bake the pizzas while in transit so as to cut delivery times and improve the finished product showing up at customers’ doors. The “baked on the way” trucks as they were called, made appearances at corporate campuses and on Jim Cramer’s “Mad Money” TV show on CNBC.

But at least five former employees cited issues with the startup’s famous trucks. The main truck, nicknamed “Martha,” cost the startup over $1 million, but lacked the technology to adequately bake pizzas, sources said.

Photos of the trucks’ interiors provided to Business Insider show broken glass, insufficient refrigeration, and pizza boxes stacked on the floor. Delivery and catering crews often resorted to cooking pizzas at the startup’s Mountain View warehouse and delivering orders in personal cars, sources said. But the quality assurance team was personally directed by Garden to keep all 11 trucks in service “until something major happened,” one source said.

Another former employee recounted a disastrous catering experience in which the team had to resort to cooking the pizzas in a toaster oven for over 100 people when the ovens in two separate trucks didn’t work.

“It was all really reckless,” one of the former employees said. This employee left Zume after voicing concerns with the machinery, specifically the metal shavings and malfunctioning blade, to Garden, and threatened to alert the FTC to misleading advertising since Zume was still marketing the pizzas as baked on the way.

So as we can see, what Zume Pizza was is essentially a company that was hilariously terrible at making pizzas, disguised as a tech company — just like WeWork was a company that was actually pretty bad at renting out office space compared to its competitors, also disguised as a tech company. The robots added absolutely no efficiency — the company delivered fewer pizzas and at a worse rate than an ordinary restaurant would, plus they had metal shavings in them. The promises that the company made were always misleading and even its gimmick, cooking pizzas on-the-go in a million-dollar souped-up van, did not work.

And yet the valuations for this company were simply insane. Softbank’s initial investment round of $375m gave the company a nominal valuation of around $2bn, even as it was losing around $50m a year. This, in spite of the fact that the entire pizza-delivery industry is only worth $10bn — and at this stage, Zume basically just had a single van that did not actually work at cooking pizzas, a restaurant with a few robots that could half-heartedly assemble pizzas with metal shavings in them, and a whole of other failed projects.

In November 2019 — after all of these problems should have been increasingly clear to any investor — there were reports that Softbank was actually considering ploughing even more money into Zume at an even higher valuation.

Perhaps mercifully, whatever these negotations were appear to have fallen through, and Zume Pizza was forced to “pivot” in January 2020. This meant laying off more than half of its workforce, and changing the business plan from pizza delivery to providing packaging and “automated delivery services” to other pizza companies.

The lasting legacy of Zume Pizza — the single invention that the company has made that it now feels might someday be profitable — is a round pizza box. They say that it’s fully compostable. According to Softbank, this company was once worth around $2bn.

So I think it’s worth saying, to people who — like me, some years ago, to be fair — only casually followed the technology industry. You will often see tech startups with ideas like this one coming up with some viral, gee-whiz viral video. They get some nice news coverage, and the promise that soon this business model might be coming to a town near you! And then maybe three or four years later you might wonder — say, whatever happened to that company? Whatever happened to that viral video? And the reality is, 90% of the time, the company has folded or pivoted to something else because it massively over-promised with technology that wasn’t ready, or a business model that simply never made sense even if the technology did work. Somehow, though, at least until COVID-19 which I’m hoping finally puts a cap on this irrational exuberance, everyone just moved on with their lives and got taken in by the next viral startup video. Or, at least, investors at Softbank did.

I don’t want to go into too much depth about some of Softbank’s other investments in this episode, but I think it’s worth listing them briefly. There’s SenseTime, a Chinese AI startup that focuses on facial recognition technology, which is able to develop its algorithms more quickly and more easily with access to large datasets from the Chinese government. This has now been blacklisted in the US for its likely uses in surveillance systems to persecute the Uyghur minority in China.


There was $1bn in View, a startup that is going for windows that can tint automatically. This is actually not a terrible idea when it comes to energy efficiency, but the fact that they have been hawking the same product since 2007 might make you worry slightly.

There’s Katerra, a construction company that sold itself as a tech company. They aim to revolutionise the $12trn construction industry using “technology and automation.” Although employees at the company don’t actually see it as a tech company, but instead as a “developer on steroids”, according to one of the few who spoke to Cretech who hadn’t signed an NDA.

One of the things that raised eyebrows when WeWork went public was that the founder, Adam Neumann, owned several of the properties that WeWork itself was renting. In other words, the founder of the business was essentially, via this arrangement, pumping Softbank’s investment VC money directly into his own pockets. And Katerra has some similarly interesting financial arrangments. One of its co-founders owns a big Private Equity firm that was Katerra’s sole customer in its early years and provided it with a dozen projects to work on. Which is all rather cozy when you can make your business look better by employing yourself.

Cretech’s article goes on to describe how, once again, Katerra’s claims to be a tech company are fairly flimsy at best.

Katerra has said its mission is to deliver construction projects that are “better, faster and cheaper.” Through tech, the company says, it can streamline the antiquated processes that plague the construction industry.

Such claims are a hallmark of a generation of startups that aim to disrupt their respective industries by replacing byzantine systems with tech. Katerra says it has applied for almost 60 patents, the bulk of which have been filed in the past 18 months. A handful of those have been approved, though they relate to electrical systems and mass timber production.

But like other firms in the SoftBank portfolio, Katerra has faced questions about whether its tech-enabled business model is deserving of the firm’s massive valuation.

Its decision to close the Phoenix factory suggests a failed attempt to achieve this goal. When it opened in 2017 as the company’s first facility, its primary purpose was to manufacture prefabricated components — including wall and floor panels — that could be shipped to construction projects. A series of high-tech machines would assemble the components, which were tracked with an RFID chip system.

But soon enough, problems surfaced with some materials made in Phoenix. At its first project in Spokane, Washington, known as Riverhouse, wall panels made of timber arrived at the job site warped and unusable, according to two people familiar with the matter. One person involved in the project said this was because the timber, sourced from the wet climate of Washington, was trucked to Phoenix, where it sat in searing heat, and became warped.

“No one understood from the beginning why we had the Phoenix factory,” said one former employee.”

That factory in Pheonix, which manufactured these prefabricated components, has since shut down. So the innovation that turns Katerra into a tech company is essentially — using wood, and prefabricating your own components in your own factory. But this is actually not uncommon for several different companies to do, having this kind of vertically integrated supply chain. So while Katerra has not had the same sort of disastrous fall that other Softbank portfolio companies have seen, it’s still tech and finance people who have spun the wheel of fortune and decided to “disrupt” construction. And you wonder how well a traditional construction company would be doing if they’d had the same levels of funding that Softbank gave Katerra — nearly a billion dollars.

There’s Compass. This is a real-estate brokerage that, again, poses as a tech company. So again, what do we actually have here? A couple of apps — one that lets you search for homes: “ In the app, buyers and sellers can search by standard things like neighborhood, number of bedrooms, price range, and so on. But they can also look at more advanced metrics, like year-over-year analysis of median price per square foot, days on the market, and negotiability.”

And another arm of the business that employs real-estate agents as “independent contractors” rather than direct employees of Compass.

Once again, as you read through the literature, it’s extremely hard to understand what the actual “tech innovation” is here. Is it a search engine for homes? But we know that there are lots of other companies that provide those services already. Is it an app to hook up real estate agents with home-buyers? You can say “machine learning” and “data-driven” all you want, but if there’s no particular reason to actually use this over the many other services that do the same thing, you have to question whether there’s actually any secret sauce here at all, or if it’s just posing as a Silicon Valley startup to drive up the valuation of the company, without any clear idea about how it wants to use any of the technologies that are vaguely alluded to.

Softbank pumped in $400m and the company overall was valued at $6.4bn. This is what an article in Barrons had to say about Compass:

“Some former Compass employees say that the company’s technology falls short of being a disruptive force in the industry or providing a significant advantage to agents. Former Compass employees and real estate professionals with knowledge of the company’s operations told Barron’s that Compass often struggles to get its own employees to use its technology.

A Compass spokesman declined to answer questions about the company’s tech platform or the productivity of its agents. As a private company, Compass isn’t required to disclose financials, and it declined requests to make financial data available.”

Sadly, regardless of how good the real estate business might be in ordinary times, it’s something else that is going to be badly impacted by COVID-19.

The final company I want to mention, just for how bizarre it was, is Brandless. Brandless was a retail company that wanted to sell products in a minimalist packaging. The spin here is that by paying for branded products, you’re paying extra for people who design the packaging, create the brand, etc. Instead, they would cut out what they called the “Brand Tax” and pass the savings on to you. Quoting from the company’s own website:

Brandless provides everyday products that are better for you, better for your family, and better for the world” while eliminating the BrandTax™ (unnecessary, hidden costs you pay with other brands).

Now you might very well think that creating a line of distinctive products with their own, minimalist packaging, called “Brandless”, that claim to be better for you and the world than other brands of product… actually involves making a brand. And you would be right, because like any other brand, they trademarked their packaging, which was co-designed with a New York Ad Agency, and they refer to themselves as a brand in their own internal literature. It’s literally just that their brand is claiming not to have a brand. In 2019, people were claiming this company was going to compete with Amazon.

And you might also question how a Silicon Valley startup is really going to be able to sell, basically, value food… any cheaper than a gigantic conglomerate like Walmart, WholeFoods, or Amazon… and you’d be right! After taking $100m of Softbank money, the startup, which only existed for three years, shut down at the start of 2020.

Is this really the Vision Fund dream? $100m for a company that just sells food with plain packaging? Is that the kind of innovation that is really going to disrupt retail? Is all you need to do to spout meaningless tech buzzwords to part these investors from their money, all the while burning through eye-watering sums of cash? Apparently yes.

So much, then, for the “wacky” arm of the Softbank Vision Fund’s terrible investments. The companies we’ve dealt with in this episode were mostly either just plain bad ideas, or else companies that are claiming to have some techno-magic USP that they don’t really have to justify the bloated valuations that come along with identifying as a tech startup and not as what you actually are, which is some other kind of enterprise.
But these fripperies are not actually the most dangerous things that the Softbank Vision Fund has invested in. Because actually, part of what Softbank has been propping up has been some financial technology, or fintech companies, that might actually be genuinely dangerous, fraudulent, or Ponzi-scheme-esque enterprises. I think — bearing in mind that I am not an economist — that some of what these companies have been doing is setting up the conditions for a new financial crisis, much like the one that we saw in 2008–9. By posing as technology companies, they have avoided some of the regulations that might have otherwise applied to banks and other lenders. And one of these companies has already been involved in what appears to be a billion-dollar fraud.

All that — and more — next time on the penultimate episode of Softbank’s Blurry Vision.

Softbank’s Blurry Vision, Part V: The Snake That Eats Its Own Tail, and the Next Financial Crisis

In the last few episodes, we’ve talked about Softbank’s $100bn Vision Fund, which has ploughed money into various different companies. A few of them have a decent business model, and one or two have been very successful. But most of them are essentially bad ideas, or poor business models, or simply bad businesses, that are massively overinflated in value by making vague claims at being technology companies, despite the fact that there’s really not a great deal of “technology” involved in what they do.

So far, so Emperor’s New Clothes. But in a couple of the fintech offerings that Softbank has invested in — which, again, have some interesting business models — there is a hint of something much darker.

The first of these companies I want to discuss is a company called Greensill Capital. I need to acknowledge that I first became aware of Greensill Capital when the amazing Trashfuture podcast — which has been mocking the wacky startups that Softbank has been pouring venture capital into for years, as part of their central thesis. That thesis, by the way, increasingly being proved correct is that a large part of the tech industry is essentially just people either selling stupid ideas at inflated valuations using technology buzzwords like “machine learning”… or that the actual innovation is not any technology, but the ability to circumvent regulations and cut corners when it comes to employees’ rights, such as Uber’s model of having “independent contractors” instead of employees. And Trashfuture also owe some of what they found out to reporting that’s originally in the Financial Times. So, again, none of this is a new story or breaking news. I’m just bringing it to you because I’ve been obsessed about this stuff for a while, thanks to the reporting of others!

So, to Greensill, who have received $1.9bn in funding from Softbank. What do Greensill do? Here’s the quote from their website.

“Greensill is the market-leading provider of working capital finance for businesses and people globally. We unlock capital so the world can put it to work.

Founded in 2011 by Lex Greensill, the company is headquartered in London with offices in New York, Chicago, Miami, Frankfurt, Bremen, and Sydney. Greensill provides Supply Chain Finance to customers around the world. The company has extended more than $150bn of financing to 8m-plus customers and suppliers in more than 175 countries in the nine years since inception.

We own Greensill Bank and created the world’s first Supply Chain Finance fund. No other bank or financial services company has our passion and expertise.

Before forming Greensill, Lex established and led Morgan Stanley’s global supply chain finance team and successfully developed supply chain finance practices for Citibank. He was awarded the CBE for Services to the British Economy in Queen Elizabeth II’s 2017 Birthday Honours.”

After reading this, you’re probably none the wiser. What is “supply chain finance”? What are “services to the economy”? Greensill, by the way, is pretty tight with the British Government — it’s involved with some contracts that are associated with the NHS, as well as the CBE for its founder, and one of the advisers on its board of directors is former British Prime Minister David Cameron.

What, exactly, does Greensill do — why is it a tech company, why are Softbank involved, and if its job is extending lines of credit to people or “providing finance” to businesses and people, what makes it any different from a bank?

These are all good questions. To get into a little more detail, then, we need to explain what “supply chain financing” is, and what “reverse factoring” is. Obviously all of this stuff is wrapped up in as many confusing buzzwords as possible, essentially to conceal what’s really going on; the more buzzwords, the better, as we have found with many of the other companies that Softbank has invested in in the past.

Imagine you’re a company that manufactures something — pencils, say. In its simplest terms, you have a balance sheet. You have a supply chain — you need to buy the wood and graphite to make the pencils! And you also have people that you supply to, who are ordering your pencils.

Now you have some debts that you owe — for buying the raw materials, and some debts that are owed to you — for selling the pencils, and all of these need to be paid on different dates.

Let’s say that I’m owed $100 by another company that has bought my pencils, and they are going to pay up two months from now. What supply chain financiers allow you to do is to get paid now, but in exchange for less money. They are intermediaries who might come in and say: “okay — that person owes you $100. I’m going to buy that money owed off you. I’ll pay $98 upfront. When they pay that money, they’ll pay me the $100 instead.” Greensill therefore makes that $2 in profit, in exchange for waiting for the money to be fully matured.

You can think of supply chain finance, then, as allowing you to “cash in” debts that are owed to you earlier, for less money. And you can see the advantages for a particular company to be involved in supply chain financing. If you’re owed money, and your rent is due tomorrow, then someone who can come along and give you the money when you need it is clearly going to be helpful for your business is good. Maybe you avoid having to take out a loan just to pay rent, or maybe you avoid even being unable to pay that money, or you avoid some interest payments, or whatever.

There is nothing inherently wrong with this as a financial service. And, indeed, if you listen to Lex Greensill he’s doing a great public service. He is keen to emphasise how he was inspired to open this company based on seeing his father’s farm and other similar agricultural businesses struggle to pay their bills on time, even though they were owed money. That might seem a little bit at odds when actually this is a guy who’s worked at Citibank and JP Morgan and so on doing similar things for years — banks are not exactly charitable organisations who exist to help people pay their bills on time, even though they are notorious for selling themselves in that way.

But obviously there is an aspect to this which is exploitative, which is that effectively Greensill is making money out of businesses that have trouble paying their debts on time. They are making money out of businesses that are having persistent trouble paying their bills by skimming them off the top a little bit. If you had good reserves of money lying around, you’d be able to pay that rent regardless of whether people owe you money or not. Instead, what managing all your invoices through Greensill does — which is what they want you to do — is encourage you to get slightly milked for cash every time someone owes you money.

It can also encourage people to keep their companies going when they are technically insolvent and can’t afford to pay their bills, by essentially extending a short-term line of credit to the company, which can convert debts that are owed to it into ready cash, even if they can’t technically afford to pay for their bills. This can lead to some pretty creative accounting at these companies.

And, of course, what it further encourages people in your supply chain to do is to set silly terms about when they’re going to pay you. So what a lot of companies do is extend the terms of their invoices. In other words, they’ll say — I’ll pay you that $100 in six months, or a year, or you can take the Greensill option and get your $98 now.

This is from the Sydney Morning Herald.

“In Australia, Greensill has faced press criticism amid concerns that some of its clients, most recently CIMIC’s UGL business, have been pushing out payment terms beyond the acceptable 30 day limits and forcing any supplier who wishes to be paid earlier to pay fees or accept discounts. Greensill has many successful big-name clients including Telstra, Vodafone and Airbus. But it has also provided supply chain finance to less successful groups. One of its recently collapsed clients is scandal-plagued London-listed hospital operator NMC Health, which called in administrators last month. Another client of Greensill, controversial rent-to-buy retailer BrightHouse also entered into administration in late March. The group also provided financing to Singapore-commodities trader Agritrade which collapsed amid fraud allegations from its lenders.”

The issue is not localised to Australia, either. You may not have heard of the company Carillion — unless you’re British. This UK construction and facilities management company collapsed in 2018, somehow getting into a position where it owed £8bn. The fallout led to thousands of jobs being lost in the UK, and thousands of small businesses losing money that Carillion owed to them. Concerns about Carillion, and its extensive use of “reverse factoring” — the same financial service that was offered by Greensill — were first raised in 2015. This is from their Wikipedia page:

“The liquidation announcement had an immediate impact on 30,000 subcontractors and suppliers, Carillion employees, apprentices and pensioners, plus shareholders, lenders, joint venture partners and customers in the UK, Canada and other countries. Subcontractors were said to be vulnerable: the Specialist Engineering Contractors Group said Carillion’s failure could lead to many smaller firms going under. Up to 30,000 small businesses were reportedly owed money by Carillion,who used ‘delay tactics’ and withheld payments to suppliers, sometimes for up to 120 days.””

Yet despite the fact that Carillion was increasingly in a shaky financial position — despite the UK Government awarding over £2bn worth of contacts to Carillion over the years after these concerns were first raised in 2015 — it took another three years for it all to finally fall apart. When it did, things were so bad that the company had to be liquidated completely, rather than going into administration, which is what financially distressed companies usually do. The whole thing was a massive disaster that led to many parliamentary enquiries, and there was a general consensus that the use of “reverse factoring”, the service that Greensill is now a multi-billion dollar company for doing, was part of what allowed it to get into such an awful state.








Quoting from a recent Financial Times piece on Greensill:

“Supply-chain finance polarises opinion, however, specifically at a time of growing financial strain due to the coronavirus crisis. Proponents stated it could help reduce damage to little manufacturers using this interruption, while experts argued it can hide dangers on businesses balance sheets.

Rating agency S&P in March warned that supply-chain finance can be held concealed from people and used to mask a far more fundamental deterioration in an organization’s monetary health.

Before its demise, construction group Carillion made hefty utilization of the UK governing bodies supply-chain finance programme Mr Greensill aided develop. MPs examining the outsourcers 2018 failure stated the scheme permitted it to prop up its failing business design. Mr Greensill informed the Financial instances last week which he thinks rating agencies should think about credit given by vendors to be an economic liability. Supply-chain finance was central on revelations of vast amounts of bucks of hidden debts and potentially deceptive transactions at NMC wellness. In four months it’s tumbled from a FTSE 100 business to close collapse, in just one of the worst governance scandals to hit the London stock market.”

So Greensill can allow companies that are effectively zombified and incapable of paying their operating expenditures to continue going for longer than they should, by cashing in the invoices owed to them early. Greensill’s business model can encourage people to stiff others in their supply chain by pushing out the terms of their payment for months, forcing people to get less money than they are owed if they want to have it in any kind of reasonable time. And obviously, in a world like the one we’re living in now — the post COVID-19 world — Greensill essentially leeches on to companies that are struggling to pay their bills, companies which are forced to accept $98 today rather than $100 in a few weeks and to go bust — so it is intrinsically quite exploitative, like a payday loan company, because the only reason you are accepting their terms is because you don’t feel like you have a choice.

But it may well get quite a lot worse than this. Because you might be wondering — how is it that Greensill has the money to pay these people in the supply chain? Greensill is almost behaving like a bank in that it is issuing loans to people, but it’s not actually a bank. Banks, under regulations like the Dodd-Frank act, have to have a certain amount of actual cash in reserve — they can’t just issue infinite loans to people without having assets to necessarily back them up. This is aimed to prevent the kind of financial shenanigans that caused the global financial crisis in 2009. But by branding itself as a financial tech company, Greensill is avoiding being beholden to a lot of these regulations — and the companies we mentioned above that might be mad about how this is working can’t pursue Greensill through the same channels, because it is not a bank.

But how does Greensill get ahold of its money? Here’s where the real concern can lie. Remember that Greensill is buying up people’s debts so that $100 you owe me is now owed to Greensill instead. But Greensill itself doesn’t keep ahold of those debts. Instead, it turns them into a financial instrument — a bond. A bond is basically just an asset that will pay out at a certain interest rate after a certain time. So you can see how a loan can be turned into a bond.




You owe me $100. Greensill pays me $98 and I transfer the debt to them. Greensill turns that debt into a bond, which it sells to someone else for $98. When the money is paid, Greensill gets $100. They pay the bond-holder $99 now that the bond has matured, and keep the $1 for themselves. In other words, providing it all works out, everyone kind of wins. It’s the arbitrage between who wants, or needs, money now and who is happy to take it later.

But by turning supply chain finances into bonds, these debts become assets. There are a couple of problems here. One of them relates to the problems we’ve identified earlier with Greensill.

Because the problem is that you can potentially change debts you owe in your supply chain into assets. You can buy bonds that are effectively debts you owe to people in your supply chain. So, via a kind of circular financing, in a weird, loose, incestuous way with Greensill as the intermediaries, you can owe yourself money. And this makes your balance sheet look a lot better, because things that would normally be listed as debts that you owe are actually now listed as assets as well, on both sides of the balance sheet, making the company look better than it really is.

What kind of company would have that sort of arrangement? Well, allegedly, and to make it clear this is all alleged at the moment, but according to reporting from the Financial Times, this is exactly what Softbank has been doing. This is all from the article:

“Softbank has reportedly invested over $500m (£399m) into Credit Suisse investment funds that made large bets on startups backed by the Japanese tech investment giant’s Vision Fund. Softbank invested in the Swiss lender’s $7.5bn range of supply-chain finance funds, the Financial Times reported, citing three people familiar with the matter. The funds are marketed by Credit Suisse to professional investors. Marketing documents sent to investors show that the funds have increased their exposure to several startups backed by Softbank’s $100bn Vision Fund over the past year, according to the FT. Four of the 10 largest investments in Credit Suisse’s main supply-chain finance fund were Vision Fund companies at the end of March, the paper reported, accounting for 15 per cent of its $5.2bn assets. These included hotel chain Oyo — which has started to retreat from a rapid global expansion as coronavirus hammers the hospitality industry — and car subscription startup Fair, which cut 40 per cent of its staff last year. Fair was also the second-largest investment in Credit Suisse’s “high income” supply-chain finance fund at the end of last year, the FT reported.

Reports of the investment come at a challenging time for Softbank, which warned last month that it may not pay a dividend for the first time since its listing in 1994 after an $18bn hit from the Vision Fund dragged the conglomerate to a record loss. Greensill Capital, a Vision Fund-backed firm specialising in supply-chain finance, selects all the assets that go into the Credit Suisse supply-chain finance funds under an arrangement dating back to 2017, according to the FT.”

So you can see how this arrangement is all really, really… frankly, suspicious. Softbank pumps money into Greensill. Greensill provides a financial service that lets Softbank companies and Softbank startups turn their debts into bonds. And then Softbank companies buy these bonds. All the while, the companies are making themselves look like better investments by turning the debts they owe into assets… but the whole thing is circular, a serpent eating its own tail. It appears as if Softbank is involved at every stage of the transaction here, facilitating a weird circular financing where Softbank companies pump money through themselves in an endless cycle.

If this was just limited to Softbank, then obviously it would be a big red flag for anyone invested in that company, or any of the startups that it backs. But there is the potential that this thing is going to go even wider than Softbank and its $100bn Vision Fund.

Let’s think back to 2008–9. Why is it that a bunch of people not being able to pay their mortgages in the US resulted in a global financial meltdown and millions of people around the world being thrown out of work?

The reason was that these people failing to pay their mortgages didn’t just affect the mortgage companies they owed the money to. Instead, these mortgages were bundled up into financial instruments called collateralized debt obligations, or CDOs. You’ll remember all of this from The Big Short. These financial instruments had become so complicated, so byzantine, and so over-leveraged that people essentially had no idea what they were buying when they bought these bonds — they just assumed that because the mortgage industry was so reliable that the bonds were really worth exactly what they were supposed to be worth.

When the music stopped, though, no one wanted to be holding any of these CDOs. No one knew how much exposure they had to bad debts that were never going to be paid. In other words, turning dubious debts of unknown origin into assets that you package together and sell works fine… until there’s some kind of crisis. Then, suddenly, people all want to sell at once; people have no faith that the bonds are really worth what they claim to be worth; and, all of a sudden, the whole financial industry is in total meltdown, banks are on the verge of collapsing, and so on.

That’s what happened in 2008–9. And of course, people did then regulate the banks — although certain politicians have tried to repeal even these regulations. But, of course, if you identify as a tech company instead of a bank, you aren’t going to be bound by the same regulations. We all remember what happened in 2008–9. The banks were seen as too big to fail, and to stop the system from collapsing, the central banks printed money in the form of quantiative easing, and bought lots of bonds to support these banks and companies. Money that was supposed to be a loan, but which doesn’t seem to have ever been paid back at any point.

You can argue that the fact that no bankers went to prison or really faced any consequences at all for the failures in 2008 would encourage people to make similar decisions in the future. And, well, unfortunately, it seems like this has happened. In a recent article in The Atlantic, called, ominously enough, The Looming Bank Collapse, the situation is described.

“To prevent the next crisis, Congress in 2010 passed the Dodd-Frank Act. Under the new rules, banks were supposed to borrow less, make fewer long-shot bets, and be more transparent about their holdings. The Federal Reserve began conducting “stress tests” to keep the banks in line. Congress also tried to reform the credit-rating agencies, which were widely blamed for enabling the meltdown by giving high marks to dubious CDOs, many of which were larded with subprime loans given to unqualified borrowers. Over the course of the crisis, more than 13,000 CDO investments that were rated AAA — the highest possible rating — defaulted.

The reforms were well intentioned, but, as we’ll see, they haven’t kept the banks from falling back into old, bad habits. After the housing crisis, subprime CDOs naturally fell out of favor. Demand shifted to a similar — and similarly risky — instrument, one that even has a similar name: the CLO, or collateralized loan obligation. A CLO walks and talks like a CDO, but in place of loans made to home buyers are loans made to businesses — specifically, troubled businesses. CLOs bundle together so-called leveraged loans, the subprime mortgages of the corporate world. These are loans made to companies that have maxed out their borrowing and can no longer sell bonds directly to investors or qualify for a traditional bank loan. There are more than $1 trillion worth of leveraged loans currently outstanding. The majority are held in CLOs.

I was part of the group that structured and sold CDOs and CLOs at Morgan Stanley in the 1990s. The two securities are remarkably alike. Like a CDO, a CLO has multiple layers, which are sold separately. The bottom layer is the riskiest, the top the safest. If just a few of the loans in a CLO default, the bottom layer will suffer a loss and the other layers will remain safe. If the defaults increase, the bottom layer will lose even more, and the pain will start to work its way up the layers. The top layer, however, remains protected: It loses money only after the lower layers have been wiped out.

Unless you work in finance, you probably haven’t heard of CLOs, but according to many estimates, the CLO market is bigger than the subprime-mortgage CDO market was in its heyday. The Bank for International Settlements, which helps central banks pursue financial stability, has estimated the overall size of the CDO market in 2007 at $640 billion; it estimated the overall size of the CLO market in 2018 at $750 billion. More than $130 billion worth of CLOs have been created since then, some even in recent months. Just as easy mortgages fueled economic growth in the 2000s, cheap corporate debt has done so in the past decade, and many companies have binged on it.

Despite their obvious resemblance to the villain of the last crash, CLOs have been praised by Federal Reserve Chair Jerome Powell and Treasury Secretary Steven Mnuchin for moving the risk of leveraged loans outside the banking system. Like former Fed Chair Alan Greenspan, who downplayed the risks posed by subprime mortgages, Powell and Mnuchin have downplayed any trouble CLOs could pose for banks, arguing that the risk is contained within the CLOs themselves.

These sanguine views are hard to square with reality. The Bank for International Settlements estimates that, across the globe, banks held at least $250 billion worth of CLOs at the end of 2018. Last July, one month after Powell declared in a press conference that “the risk isn’t in the banks,” two economists from the Federal Reserve reported that U.S. depository institutions and their holding companies owned more than $110 billion worth of CLOs issued out of the Cayman Islands alone. A more complete picture is hard to come by, in part because banks have been inconsistent about reporting their CLO holdings. The Financial Stability Board, which monitors the global financial system, warned in December that 14 percent of CLOs — more than $100 billion worth — are unaccounted for.”

The whole article is worth a read. It’s pretty chilling stuff. But essentially, what we are seeing is that precisely the same conditions that prevailed before the financial crisis — greed, greed, and more greed, obscured by financial jargon, a system that is corrupt to its core which is totally invested in pretending that everything’s fine for as long as possible, safe in the knowledge that taxpayers will foot the bill, because the risks for these activities always seem to end up being paid for by you and me in some way.

And this packaging of loans and debts to distressed companies into bonds, and assets, and funds made up of thousands of different bonds in a less than transparent way… this financialisation of everything, turning it all into financial instruments to be traded, bought, and sold, repackaged, and reissued… this has all been aided and abetted by Softbank through Greensill, and Softbank’s Vision Fund which is doing it, frankly, for its own purposes.

So what’s the idea here? What’s the innovation? What’s the technology breakthrough? All very good questions. Because it seems like the innovation is basically just to relabel all of the extremely dodgy practices that led to the last financial crisis and major global recession. And while these kind of things might fly when the market is bullish and everyone’s feeling nice and optimistic, the consequences when it all comes crashing down are much worse. The COVID-19 pandemic and the global recession is obviously going to mean that a lot of companies are not going to be able to pay back their loans. They are going to be defaulting left right and centre. Supply chains will collapse. And, because suddenly everyone is horribly over-exposed to loans that they might not know anything about, banks may very well collapse as well. The consequences of irrational exuberance like this are always, always a worse crash when the crash happens. And I’m afraid that the suspended animation that the economy has been put in by central banks and world governments, in this “emergency response” mode to the COVID-19 pandemic, is not going to last… and when the music stops, this will all start to unravel. Heck, the fact that I’m telling you about it as a lowly student who knows very little of the financial sector is probably enough to tell you that it’s already unravelling, and fast. I’m very afraid that we will all suffer as a consequence.























Softbank’s Blurry Vision, Part VII: COVID-19 and What We Lost in the Hype


Over the last few episodes, I’ve outlined the story so far of the Softbank Vision Fund — $100bn which was raised, supposedly to revolutionise the world and accelerate the deployment of all this glorious new technology. And I’ve discussed how a lot of this money has been ploughed into unsustainable or actively harmful business models; into silly ideas that never had a prayer of working; and into “fintech” companies that either have, or enable, sketchy accounting practices that could well be inflating yet another financial bubble. I have discussed how the same common features — avoiding regulations, aggressively driving competitors out of business, and claiming to be a technology company that can grow like a technology company — flow across many of the Vision Fund’s major investments.

I hope I’ve convinced you that any company the Vision Fund invests in is worth a second look, and that it’s always worth asking what the real innovative technology behind a “tech company” actually is. I hope I’ve convinced you that we’re witnessing a big bubble bursting right now, in the midst of the pandemic, as these companies continue to fail. And I hope I’ve convinced you that, just because someone has billions of dollars and has succeeded in the past, it certainly doesn’t prevent them from blowing up bubbles and wasting all of that money.

In some ways, the game now appears to be up for the Softbank Vision Fund. In the last quarter of 2019, the profits made by Softbank the company were wiped out by losses at the Vision Fund, and naturally the Vision Fund has already lost a great deal in the COVID-19 pandemic as many of its businesses fold. While there were plans, as late as Summer 2019, to create a second round of Softbank Vision Funding at another $100bn, those plans look all but certain to be shelved in light of the global technological depression. Softbank, the company, will probably survive in some form or another — and, perhaps in 10–15 years, when the market reaches the height of irrational exuberance again, Masoyoshi Son might get involved once more in whatever the bubble is for that particular generation of investors… but, at least for this round, the Vision Fund is winding down, its bets made, and seems more likely to be remembered as a bizarre experiment at the height of this exuberance around tech that has ultimately failed… rather than the start of a thousand-year plan for Softbank to achieve world domination.

But beyond this just being a story of wild excess and terrible decision-making in Silicon Valley — beyond us just getting to laugh about robots that can’t make pizza and erratic CEOs with egos the size of a planet — why should you care about this? After all, if Softbank, and the Saudis, and the UAE want to blow a whole load of money on these bad bets, what does it mean to me?

First off, I want to point out that irrational exuberance and bubbles, like the one Softbank has been blowing up, hurt a lot of people when they pop. Now you might not mind if Masoyoshi Son or the Saudis end up losing a fortune, and I’m kinda inclined to agree with you there. But it inflates all of this unsustainable economic activity. Somehow, the CEOs and the founders seem to escape, if not with billions, then at least with millions. But obviously the employees at these companies, who sometimes get paid in stock options that end up being worthless, lose out. They’ve then put a great deal of their careers into something that was always unsustainable and unlikely to work out. When WeWork collapsed, it was so insolvent that it couldn’t even make people redundant because it couldn’t pay them any severance money — it had to wait on a cash bailout before it could even fire people. All this, while it was still planning to give founder Adam Neumann nearly $2bn.

We know that the Softbank Vision Fund encourages Softbank employees to invest in it. If it goes under, should the employees really pay for Masoyoshi Son’s bad approach to investing?

Another major reason, of course, depends on how cynical you feel about these bubble tactics. After all, the aim of these private companies backed by Softbank is to eventually go public. When Softbank pumps money into these companies, it is giving them a nominal value; it is associating them with this tech hype. If a company gets a billion in cash from Softbank, then it must be worth $10bn, or $20bn — not because of any underlying business case for its technology, but because it has attracted so much cash. At its very worst, this is a confidence game, aimed to inflate the value of these companies so that investors will buy Softbank’s stake back at a profit.

We have already talked about how the Vision Fund is backed by a great deal of debt, as many of Softbank’s investments are — structured like bonds that promise to pay a profit, but are ultimately debt. So, through a certain lens, Softbank’s main business model is just to pump billions of borrowed money into private companies, swelling their valuation — then sell their stock on when the company goes public so that Softbank won’t foot the bill when they collapse. In the case of Wirecard, this was done despite allegations of fraud. It would’ve worked in the case of WeWork if its IPO filing wasn’t so absurd that most investors were turned off straight away.

There will always be people in the investment world whose model is, at its core, something like pump-and-dump: they have no interest in the underlying value of what the company is doing, but instead in making sure they can sell on their stake in it at a profit when the time comes, and get out just before the crash. If you’re feeling charitable, Son is a dreamer who doesn’t mind making losing bets. If you’re feeling uncharitable, then this is the real game here. And it’s the kind of game that can work very well for a few people until the music stops.

A third major reason, which we’ve covered extensively in the episodes on Uber and WeWork in particular, is how the business models for these companies work. They claim to be innovative; they paint fanciful narratives about changing the world and indulge in techno-hype in order to inflate their share prices and attract financing. Softbank, of course, never has any motivation to deflate the hype in companies once they’ve been backed. But in reality, they offer this very predatory business model.

It’s often been commented that the first generation of very successful tech companies can offer you things for free because you are the product. Facebook and Google offer their services for free — and in exchange, profit from the information that they gather about you, in the mechanism of surveillance capitalism. They lure you in for free and you pay for it later.

But in the second generation we are also initially getting things cheaply or in a heavily subsidised way — loss-making taxi rides, loss-making office buildings — in the hope that they can make you pay for it later once a monopoly is established. Quoting now from Vice:

“A 2018 study by University of California researchers Martin Kenny and John Zysman maps out the same period and explains that the explosion of the number of start-ups, the proliferation of unicorns (start-ups valued at $1 billion or higher), and the unprofitability of a majority of unicorns when hitting the public market are a consequence of them “each trying to ignite the winner-take-all dynamics through rapid expansion characterized by breakneck and almost invariably money-losing growth, often with no discernible path to profitability.””

If this fails, Softbank can always hope to sell its stake on at a higher valuation. But when these businesses do eventually fail — and we may see this start to happen with Uber, as it’s already happened with WeWork — they leave this scorched earth in their wake. Gig-economy employees who have very few rights, end up working excessive hours for little pay, or barely any hours at all — or who end up enticed in by the promise of high wages that don’t take into account the repayments on their cars. All the while smaller, profitable businesses are driven out and eaten alive by the debt-financed VC engines like Uber and WeWork, or Oyo in the case of the hotel industry. The result is that the “independent contractors” in these companies no longer have any choice about who to work for, or what conditions they have to work under, as the traditional industry, for all of its flaws, has been destroyed. It’s quite amazing that companies can be wildly unprofitable, terrible for workers, destructive of existing industries, and bad for society in general. Again, as Vice puts it:

“We do not have companies like Uber and WeWork because they’re efficient or innovative or even because we want to, we have them because they are being subsidized by venture capital. And here’s what we have to show for it: an underclass of gig workers, increased traffic congestion and urban pollution, the global suppression of labor standards, hollowed-out public transportation and taxi businesses across the world, and the instability that will come when Uber and WeWork collapse as SoftBank and other investors get tired of losing money from these creatively unprofitable businesses.”

The net impact of many of these businesses focused on disruption as a goal is that it has made things worse for nearly everyone except that we have this totally unsustainable expectation that things should be available swiftly and on demand. In the Silicon Valley mould, then, tech isn’t about creating — it’s not about allowing people to do stuff they couldn’t already do — but coming in, gutting an industry that may have had its issues but was working before, throwing a whole bunch of people out of work. These companies are not using technology to improve lives, as we know it can do when it’s deployed well and correctly, but instead they are engines of inequality using people’s love of tech and infatuation with the future as a veneer to try to smash and grab value from various different industries that used to employ people more sustainably.

Of course, this represents a huge waste of potential. A huge, huge waste of potential. Let’s first look at the money. $100bn over the course of five years in the Softbank Vision Fund, the vast majority of which has been wasted.

UN Rapporteur on poverty, Philip Alston, pointed out that even by the paltry standards of the current definition of extreme poverty in the world, 700 million people live on less than $1.90/day. The Softbank Vision Fund alone could have lifted nearly 30 million people out of poverty across the five years that it has existed.

The International Food Policy Research Institute came up with a series of estimates for how much funding would be required to end world hunger. This is obviously extremely difficult and subtle to model, and you end up with a big range of estimates depending on how you define world hunger and also how you define eradication. But one of their central estimates was that you could make a very serious dent in the problem for $11bn a year. The Softbank Vision Fund money, over the last five years, could have prevented millions from going hungry or starving to death.

You might say: “Okay, but that’s charity. You’re not going to be able to persuade the Saudis and Softbank to invest in something that won’t provide any kind of economic return.” And yet the IFPRI also points out that, if you were to eradicate world hunger, it would more than pay for itself: the estimated economic benefit from doing so would be as much as $276bn. They note, drily, that expressing it in these terms “is useful for advocacy and policy prioritisation”… in other words, if you’re living in a society where people can only make decisions based on comparing numbers expressed in dollars, it might be worth pointing out that alleviating world hunger is incredibly valuable. Surprise — people who aren’t malnourished lead healthier lives and they’re also more productive workers, if that’s all you care about, but, of course, also, there’s substantially less suffering involved.


Compare the $100bn that Softbank gathered up for its Vision Fund to the investment in other major organisations. The World Health Organisation, much maligned lately, our first line of defence against global pandemics: around $3bn a year. The Softbank Vision Fund could fully fund the WHO for 30 years. For $5–6bn — less than was ploughed into WeWork — you could eradicate polio, according to the WHO’s plan of action on the same from 2013.

Take any issue that you care about. The world’s governments combined, according to the IEA, invest around $3bn a year in renewable energy research and development. $4bn a year in nuclear. $1bn a year in energy storage. Since part of the motivation for the Saudis and the UAE to invest in Softbank’s Vision Fund is supposedly a desire to diversify away from oil and fossil fuels, surely it would make sense to be invested in their replacements?

Of course, it’s not like there’s a lack of solar resource in Saudi Arabia. You would need to be incredibly ambitious to do it, but if you were to blanket the desert in solar panels, it could generate energy equivalent to the entire proven oil reserves of Saudi Arabia in just two years. Yet the Saudis themselves have been on and off about the possibilities of solar power — promising much, setting grandiose and ambitious targets, but ultimately seldom delivering very much.

In fact, Softbank has been involved in one of these grandiose promises that didn’t amount to very much — and I reported on it for Singularity Hub. Back in March 2018, they announced, to great fanfare, what would have been the largest solar project in history — 200GW of solar panels in the desert. This was a collaboration between Softbank and Saudi Arabia. Mohammed Bin Salman, the Saudi leader, described it as “a huge step in human history.” The project was due to start with a $1bn investment from Softbank and would end as a $200bn, 200GW behemoth by 2030. By comparison, at the time, the installed solar capacity across the entire world was only 300GW. So it would’ve constituted 2/3rds of the entire solar capacity across the world, in one single megaproject.

At the time, there was some skepticism given how light on actual details the announcement of the project was, and the fact that the only evidence for its existence was a “memorandum of understanding”, which essentially doesn’t guarantee anything at all. Not to mention the fact that both the Saudi government and Softbank had previously bombastically promised to invest massively in solar power, before quietly cancelling the projects. The skeptics proved to be right, and, by October, the project had been quietly shelved, without any apparent work happening on it at all. You have to question whether the whole thing wasn’t just an exercise in reputation-laundering, given how little serious effort was actually invested in it before the announcement was made. After all, it’s not a bad strategy to loudly announce that you are going to do something bold, visionary, or charitable… and then quietly row back on the promise later, knowing that this will create far fewer bombastic or dramatic headlines. I regret giving the announced project the credibility I did in my report back in 2018 (even though I did finish on a very much “wait-and-see-what-happens” note.)

The Wall Street Journal in 2019 reported that of the 220GW that Softbank had suggested could be built across India and Saudi Arabia, just 3GW of deals actually existed — maybe just 1.2% — and substantially less was under construction.

And as far as the Softbank Vision Fund goes, the only energy-related investment I could find was $110m in energy storage company Energy Vault. In other words, despite all of the rhetoric and bombast about investing in the future of the world, and despite the willingness to appear to invest in solar, less than 0.1% of the Vision Fund has gone towards the energy transition. They spent multiple times that on robots delivering pizza.



This is part of a broader trend, and in some ways, it’s understandable, because it is a different kind of investment. If you spend money on a solar farm, you might well extract a nice little profit out of it. Within a few years, the electricity you sell will have paid back the initial costs of construction, and from there, you might well earn a few % a year on your project — a tidy margin. But you won’t get the sort of returns that a lot of venture capitalists are looking for. You are unlikely, for example, to invest in a solar farm that will suddenly be worth ten times, or a hundred times, more than your initial investment: that’s the nature of the infrastructure project. If you choose, design + operate it well you might get x% a year return, rather than a doubling or tripling of your initial investment.

So you can argue that it’s inevitable that an entity like the Vision Fund, which is essentially a gambling operation that is hoping to win big on a few of its bets, is not really going to be geared towards more steady and sensible investments, and will instead be far more liable to plough its money into ideas that promise explosive growth, even if they do turn out to fail in the end. You can criticise doing this as a greedy or reckless way of spending money — and wasteful, or at the very least irrationally exuberant, to plough millions into Zoom Pizza or WeWork based on an inspiring fifteen minute chat you had with the owner. But it is at least a strategy that explains why the Vision Fund wouldn’t invest in technology that is more obviously beneficial to humanity.

But this line of argument does not explain why the Vision Fund wouldn’t invest in early-stage companies that are developing new kinds of solar panel, or new kinds of battery. There are dozens of such projects that are found in university labs which struggle to be spun out into companies which Softbank could invest in. Imagine you have the investment in the product that turns out to be the next lithium-ion battery, or the coating that you can apply to all solar panels to make them twice as efficient: this is the sort of VC investment that genuinely can explode in value over time. Yet Softbank only invested 0.1% of its $100bn into this type of project.

And this is indeed part of a broader trend in VC. Elizabeth MacBride wrote a great article on this in MIT Tech Review back in June 2020. It’s titled: “Why venture capital doesn’t build the things we really need”, and the subtitle is: “The funding model that made Silicon Valley a global hub excels at creating a certain kind of innovation — but the pandemic has exposed its broader failures.”

In the article, MacBride notes that the whole VC industry has grown rapidly over the last few years. In 2005, VC investments in the US were $170bn, but by 2019 this had swelled to $444bn.

MacBride writes:

“Venture capitalists sell themselves as the top of the heap in Silicon Valley. They are the talent spotters, the cowboys, the risk takers; they support people willing to buck the system and, they say, deserve to be richly rewarded and lightly taxed for doing so.

The image, however, doesn’t strictly match the history of the Valley, because it was “the system” that got everything started. After Sputnik launched the space race, the federal government poured money into silicon chip companies. Historian Margaret O’Mara documents this well in her book The Code: In the early 1960s, the US government spent more on R&D than the rest of the world combined. While that fire hose of cash flowed, the first venture capitalists found many winners to bankroll.”

Of course, nearly everyone listening to this probably knows that the internet originally rose out of the US military — as DARPAnet, from the Defence Advanced Project Research Agency — and that the World Wide Web and many file transfer protocols that formed the earlier versions of the internet arose from researchers in CERN and other universities looking to exchange datasets and research with each other, long before there was a Silicon Valley that got involved in these projects.

“The link to government is still very much there in today’s technology companies. Google’s early work came out of the Clinton-era Digital Libraries project at Stanford, and the CIA was Palantir’s first customer in 2003 — and its only one until 2008.

O’Mara says there isn’t anything wrong with tech companies’ being built through US research dollars. In fact, she argues, the most important decision of that era was for the government to pour money in without exerting too much control. But, she adds, a mythology has grown up that focuses on lone heroes and rule breakers rather than the underlying reasons for a company’s or technology’s success. “Hooray for the internet that it’s still cranking,” she says. “But you did not do this by yourself.””

And this is not to say that there is no role for the private sector in innovation. Instead it’s simply to point out that a lot of what companies do — and get most of the credit for, and make profit for — is commercialising research that was originally in the public sector. Smartphones are a ubiquitous, multi-billion dollar industry. For most of their apps to function, they rely on code that was developed by universities and researchers, exchanging information over the internet which was developed by researchers, using GPS and satellites which were pioneered by the military. Even the touchscreen that phones rely on was originally invented by the Royal Radar Research Unit in the UK, which was an arm of military research.

What companies like Apple have done is essentially — and, obviously, rather well — brought together all of these technologies into something that’s incredibly commercially successful. I don’t say this to diminish the skill of what they’ve done, or even the value of the marketing that surrounds something like an iPhone. But there’s this general, extremely unfair perception that government operations are wasteful and the private sector is invariably sleek and efficient, especially when it comes to Silicon Valley and the tech sector. I hope, if nothing else, this series has convinced you that this isn’t always true.

So you have to look at the innovation landscape and understand that innovation, that the pure development of technology, has actually seldom come from venture capitalists pouring money into projects with an eye to near-term profitability. Military innovations arise because you’re trying to solve practical problems — and sure, it’s helpful if it can be done at a low cost, but the main point is trying to solve the problem. Primary research innovations arise out of curiosity, or with no conception of how they will ever make money. When Benjamin Franklin was asked about what could be done with electricity as he experimented with it, he supposedly said: “What use is a newborn babe?” And, of course, building a massive particle collider to probe particle physics struck no-one as a particularly profitable operation. But in the course of solving problems surrounding that, a great deal of the Internet and the World Wide Web was developed.

Solar panels are a great example of the relevant roles that the public and private sector have often played. The effects of solar photovoltaics were first discovered back in the 1800s by a guy called Becquerel when he was 19 and fiddling about in his father’s research laboratory. Solar panels first found their niche in satellites.

Recently, as prices have come down, we’ve seen the incredible power of market forces which have caused prices of solar panels to plummet by 90% in the course of a decade until they are now competitive with fossil fuels across most of the planet. The private sector, alongside lots of university researchers, have proved very good at making efficiency breakthroughs — once the path to profit is clear. And, indeed, this incredible research and development effort is a big part of the reason why we’re going to see fossil fuels continue to die out of our electricity supply in the next few decades.

The irony is that Solar photovoltaics, alongside other revolutionary inventions like the transistor, were worked on quite substantially at Bell Labs. This was a private company and laboratory that just has a truly stunning wealth of developments behind it — something like 12 Nobel Prizes awarded for work that took place there at one time or another. This is the sort of research ecosystem that you would presumably be interested in creating, if you were Softbank.


But, seeking near-term profits, private companies have moved away from investing in the kind of blue-sky research that was done at Bell Labs. After the 1980s when its parent company lost its monopoly, the labs have shrunk in funding and in size over recent years, with many people moving on from the company. A Nature Materials editorial back in 2006 pointed this out:

“To state that companies in the technology sector do not invest enough money in research and development would be wrong. According to their annual reports, IBM spends about 5.8 billion US dollars, Hitachi 3.4 billion, Alcatel 1.9 billion and Lucent 1.2 billion on research and development each year. The question is whether they engage enough in basic research, compared with the development of new products. Here the situation might be less encouraging. It seems that to some degree basic industrial research is declining.”


So a big part of the issue here is that there’s a misapprehension about where the real technical innovation comes from that makes entirely new industries possible. Much of the credit, and funding, goes to the private companies that can commercially exploit these inventions, and we look on them as the source of innovation — we look to them to solve problems.

When the companies are good, they can find a way to make money from fundamental research and innovations. When the companies are useless, like so many that Softbank has invested in, they just sell the story of innovation and technological supremacy to justify vast investments. And we see this in the Softbank Vision Fund, structured as it is in this debt-driven way and with the requirement to deliver profits to its investors. It’s pretending to be investing in the far-flung, sci-fi future — a thirty year business model — but it’s actually just chasing quick profits in this lopsided innovation ecosystem.

And the real issue here, as MacBride argues quite eloquently, is that venture capital has a very specific notion of the kinds of companies that can create the dramatic returns that they are looking for. They want to see something that can grow explosively, and to a lot of them, that means software and not hardware. It means “disrupting” an existing industry and expanding to steal its existing profit base, while cutting costs because you’re using an app and “machine learning” instead of human labour and physical facilities. In other words, it means looking for the next Amazon. But if you are looking for something that resembles an existing business model, you’re making the exact same mistakes that a machine learning algorithm does. These algorithms can only extract data on what they’ve already seen: they cannot actually predict; there’s no creativity there. Just as the algorithms will turn up search results of nerdy white guys when you search for “tech CEO”, so this is what the VC founders are often looking for. This is from MacBride’s article again:

“But some of the other inputs, either consciously or subconsciously, have been assumptions about the kind of person who can help generate outsize returns. The top founders “all seem to be white, male nerds who’ve dropped out of Harvard or Stanford and they absolutely have no social life,” John Doerr of Kleiner Perkins — one of the most influential investors in the Valley — noted in 2008. “So when I see that pattern coming in … it was very easy to decide to invest.””

Perhaps this kind of bias — about the expectations for what the next big tech company will be; that it will be software that can disrupt an existing industry and have a quirky, eccentric founder — is part of what led Softbank to make its big bets on companies like Uber and WeWork, which fit this profile, without really considering whether their business model was sustainable, regardless of whether or not it was actually creating some kind of social good for the general public.

And the reality is that this is a very, very limited definition of innovation and what technology can do, and what it can be. You’re geared towards creating these lost-leader companies that want to replace entire industries with apps. In some areas, you can actually make money doing so, even though you throw a lot of people out of work in the process. In other areas, the margins are too tight; expanding as quickly as the venture capitalists push you to (think of Son telling the WeWork founder that he wasn’t ambitious enough) is going to lead you to a disastrous business model, and you’ll just burn through a lot of cash and put a lot of people out of business trying to establish a monopoly. You have to sell yourself as vastly more innovative and futuristic than you actually are, because real developments take a great deal of time and engineering.

This is again from the MacBride article; she writes:

People who really study innovation systems “realize that venture capital may not be a perfect model” for all of them, says Carol Dahl, executive director of the Lemelson Foundation, which supports inventors and entrepreneurs building physical products.

“In the United States, she says, 75% of venture capital goes to software. Some 5 to 10% goes to biotech: a tiny handful of venture capitalists have mastered the longer art of building a biotech company. The other sliver goes to everything else — “transportation, sanitation, health care.” Dahl told me about a company that had developed reusable protective gear when Ebola emerged, and was now slowly ramping up production. What if it had been supported by venture funds earlier on?

That’s not going to happen, Asheem Chandna, a partner at Greylock, a leading VC firm, told me: “Money is going to flow where returns are.”

So how can that change? The government could turn on the fire hose again, restoring that huge spray of investment that got Silicon Valley started in the first place. In his book Jump-Starting America, MIT professor Jonathan Gruber found that although total US spending on R&D remains at 2.5% of GDP, the share coming from the private sector has increased to 70%, up from less than half in the early 1950s through the 1970s. Federal funding for R&D as a share of GDP is now below where it was in 1957, according to the Information Technology and Innovation Foundation (ITIF), a think tank.

In other words, the private sector, with its focus on fast profits and familiar patterns, now dominates America’s innovation spending. That, Dahl and others argue, means the biggest innovations cannot find their long paths to widespread adoption. We’ve “replaced breakthrough innovation with incremental innovation,” says Rob Atkinson, founder of the ITIF. And thanks to Silicon Valley’s excellent marketing, we mistake increments for breakthroughs.

In his book, Gruber lists three innovations that the US has given away because it didn’t have the infrastructure to bring them to market: synthetic biology, hydrogen power, and ocean exploration. In most cases, companies in other countries commercialized the research because America’s way of investing in ideas hadn’t worked.”

When we think about the great technological innovations throughout human history — harnessing the atom, the power of electricity, telecommunications, aircraft, photovoltaics, the internet — we start with some very early-stage experimentation that was never really likely to justify itself on a commercial or financial basis. All of these cases flowed either from government, from private individuals with money and time to do whatever they wanted, or from research labs at private companies that had a lot of time and money to work on new projects.

And, eventually, they flourished into entire industries. But an app is never going to become an industry. For all the talk of “disruption”, that’s all it is; trying to come along and push out someone who’s already there, rather than creating anything genuinely new and revolutionary. So the blinkered notions of what tech companies are, how they should behave and what they should look into… these notions of what counts as innovation and what makes money… are arguably holding us back and keeping a lot of this capital tied up in bets that end up being bad ideas, or don’t have any real potential to transform society in a positive way… and lots of them fail to make money, too.

And you can see it in the gap between rhetoric and reality when it comes to so many of these companies and so many of these investments — the fact that a great deal more patience is often required before anything is ready. Either the pronouncements are overambitious — and you end up with a bunch of robots making pizzas filled with metal shavings, or endless claims that exaggerate how close self-driving cars are to fruition — or the claim to be innovative is spurious or overblown, and you end up with companies that aren’t really companies, which pose as tech companies.

So we have a huge waste of potential. But, of course it’s not just the money. There is a vast waste of time and intellectual resources in these exercises. The engineers who have to spend all day building software that ends up being useless. The enthusiastic young people who move into a technology industry that can’t innovate to provide the things that people need. The rhetoric about changing the world, vs. the reality of businesses propped up with endless VC funding.

Just as the finest minds of my generation are set to work figuring out how to more effectively serve ads to particular people, because that’s what the market demands, rather than doing anything particularly important…

I remember my time at the Machine Learning conference, ICML, in 2019. The Machine Learning for Climate Change and AI For Social Good were the main events that I attended there — but, compared to the other shows on, many of which related to advertising and problems adjacent to advertising, they were in smaller rooms. The attendance was small, but enthusiastic. We can do this better — the paper that came out of that conference had hundreds of ideas for ways that actual machine learning tools might be deployed to help solve societal problems — but instead, a lot of this intellectual might is being devoted towards serving up ads because this is where the capital is.

And when the arm of your system that is supposed to be inventive, creative, risk-taking, bold, looking towards the future, spends five times as much on a robot pizza delivery system than it does on anything to do with finding alternative sources of energy that won’t run out and don’t choke the planet, you can’t tell me that the system isn’t broken. The system is broken. And it’s funnelling time and resources towards things that don’t matter when there are so many fundamental problems we have yet to solve.

This backlash against Silicon Valley has been building in the last few years. But I think the COVID-19 pandemic has really exposed it for a lot of people and accelerated the backlash against the innovators and entrepreneurs who yammer on about saving the world from superintelligent AI and colonising Mars and so on. When the crisis came, it showed up the massive flaws in our society and what we had been prioritising for so long.

What we needed was healthcare; infrastructure; pandemic preparedness; plans for robust surveillance and monitoring of the illness that could be put in place to nip it in the bud. Instead, in the US, for the first month or so, the CDC couldn’t even tell you how many tests had taken place — likely in the low hundreds. Here in the UK, the government abandoned testing in the community after the first few hundred cases had been observed.

Despite the many warnings from scientists that another coronavirus could spread into humans — just like SARS and MERS had done — the SARS vaccine research was abandoned when this disease was wiped out, because it would not have made a profit. Well, Uber and WeWork haven’t made profits either, but they somehow haven’t found it hard to come by billions of dollars. Perhaps if that SARS research had accelerated the ability to develop a SARS-COV-2 vaccine, it would currently be one of the best bets made in history, economically speaking. That would have been a daring venture, and a useful one; but, as described, the money was going into software instead.

Even in the areas that our current infrastructure of innovation should supposedly help, it failed to deliver. The apps that were going to help us track and trace contacts for the virus weren’t readied in time, or, in many nations, at all. Even simple things like the data for testing were being recorded, transmitted, and stored in inefficient and incorrect ways. In the UK, for example, we’ve simply given up on recording how many people have been tested (although not how many tests there actually are) because we can’t find out that information. The data pipelines are leaky.

So it’s not just that the current innovation ecosystem ploughs lots of money and smart people and intellectual effort and careers into bad ideas, but it’s come up empty when it comes to actually solving problems. Despite all the talk of future vision, something as eminently predictable as a pandemic left us reaching for tools to deal with it. And yes, it’s not all the fault of Silicon Valley, of course, governments have to take the blame in terms of what they’ve funded and encouraged as well: I would never deny that. It’s a society-wide problem, of which VC is just one of the most egregious aspects of it. All this without mentioning, of course, that the ventilators hurriedly manufactured by various different tech companies didn’t work, either. So you can list a thousand problems we could’ve been solving with this time, money, and effort, but which simply weren’t going to happen. That’s one reason to be mad.

I also think that we need to take into account the fact that replacing people with software, as a lot of these “disruptive” industries often aim to do, happens for its own reasons and has its own inevitable impacts. We know that this will be a way that the current tech industry seeks to scale and move into new industries, because software has this glorious property that it is very easy to scale up and replicate — you make a tweak to an app in a single office and it can be spread around the world near-instantaneously if desired, and the cost of copying a piece of software is miniscule once everyone has all of the hardware that’s required to do it. The same things can’t be said for physical infrastructure. And, of course, paying a handful of software engineers to update a programme that automatically does x is vastly less expensive than paying hundreds of people to do x.

So there is always going to be a temptation to replace people with software. But this doesn’t necessarily come down to the fact that software can do the job any better, although the focus on big data is going to emphasise the idea that it can. We know that there are lots of use cases where using software to do things is infinitely more efficient — in the same way as using robotic labour to do basic and repetitive tasks is infinitely more efficient.

But it should also be clear that we have to be very careful about assuming that a world run by algorithms and machines is inherently superior. We know that the algorithms themselves can automate and increase inequality in society — by concentrating profits for a small number of people who own the algorithm and own the data that everyone uses, sure. But also in perpetuating inequalities that already exist when they are used for things like hiring and firing, and prison recidivism, etc. which should be the domain of people.

Because ultimately, as much as we might hype artificial intelligence and machine learning, what algorithms univerally have to do is to reduce everything down to a single, fungible metric which can miss a great deal of nuance. This is how decisions on how good a person is (social credit score), how employable they are (HireVue), or how likely they are to commit crimes can be reduced to a single, impenetrable number. The algorithms are unbelievably efficient at producing these numbers and these decisions. But they can be fantastically bad at explaining how they came to these decisions — and they can also make terrible decisions which are covered up by our misplaced faith in the precision of technology and “AI” to generate insights that no human could possibly come up with. After all, it’s driven by huge amounts of data from a vast array of sensors — surely it must know what’s better for us more than we do.

And in the meantime, for better or worse, the rapid expansion of this kind of software is automating away people’s jobs, and the industries that develop this kind of technology — as much as they talk about solving great social problems, taking humans to space, etc. — seldom speak about that kind of problem. We’ve discussed this at length in our series on technology, inequality, and global catastrophic risks. We’ve talked about how support for a UBI from various tech founders might just be there as an excuse — in their eyes, they are going to automate millions of jobs out of existence, but they need people who can continue to buy and pay for their products and who aren’t going to be so penniless that they decide to vote in revolutionary politicians, and so they support a basic income as a kind of bare minimum reform… in much the same way as liberal reformers in Britain and Europe have often been motivated to provide “concessions” to what people want to avoid the spectre of communism. The irony here, of course, is that a lot of these Silicon Valley founders will push the responsibility for solving this particular problem onto the state while avoiding taxes and regulations from the state that might actually come close to enabling them to pay for it.

And I think this is broadly why I’m so critical of Softbank’s Vision Fund and this whole area of tech VC in general.


I could’ve spent a whole bunch of episodes criticising other areas where people waste money, or make money based on very dubious schemes — finance in general, the defence industry, etc.

But technology is different. It’s the modern-day secular religion — not just in the case of genuine Singularitarians who are counting down the days until the general AI they’ve been promised materialises out of a fog of algorithms and Python scripts — but for all of us. We all anticipate that positive changes in our lives are going to arise out of the technology sector. We are all bracing for the next set of great changes that we anticipate — this mass unemployment, self-driving cars, AI, robotics, whatever it might be.

And, of course, the tech sector is all too happy to encourage these rosy visions of the future; to hype up near-term space travel, martian colonisation, superintelligent AI, and so on. Much like a Softbank company and Softbank working arm-in-arm to tell us that they are revolutionising x, y, z.

They want to sell us this future, and we want to believe in it. Prior to the crash, many of us, myself included at times, gave Softbank a pass, or wrote excitedly about what the Vision Fund might be able to achieve. I wrote for many years for a website dedicated to hyping up technologies… and although my articles usually ended up with some variation of “it’s actually a lot more complicated than that…”, the headlines were often glowing enough.

We look to technology to save us from the problems that exist in our society, while we’re overlooking solutions that might be less attractive but far more important.

Look at Japan — seeking to rely on robots to get developed that will look after its aging population, rather than, for example, encouraging immigration which might reverse the trend of the aging population in that country.

Look at our approach to mental health. Depression and anxiety are hugely on the rise, especially amongst young people.

From 2005 to 2017, rates of depression have gone up by 52% in young people. Many have suggested that a large part of this is down to social media, which is crowding out face-to-face social interactions, while encouraging people to present a “perfect”, Instagram-worthy lifestyle to the outside world.

I think there’s definitely some truth to that. But I’d also note that, since I’ve started paying attention to the world around me, I’ve seen those who are nominally in charge fail time and again to tackle climate change, which is instead left as a problem for the next generations to clean up, as well as a disaster that will befall them. I’ve seen those nominally in charge fail to address the rising tide of income and wealth inequality in society and demonstrate little foresight to address the multiplying existential risks that have arisen, focusing instead on innovations that enrich a small number of tech billionaires at the expense of everyone else…


I’ve seen a financial crisis and global depression caused by greed for which few were punished; I’ve seen a recovery that focused on assets and didn’t reach everyone, and, in my country, government austerity proposed as the solution, which led to the University education our parents got for free lumbering us with mountains of debt. I’ve seen that, while my parents left that University education with worse grades, from a less prestigious institution, with dozens of job offers directly given to them, my colleagues and friends have applied for hundreds of jobs without success.

I’ve seen house prices rise to become utterly unaffordable for anyone in my generation; I’ve seen the statistics that show we will be the first generation to be worse off than our parents in centuries. I’ve seen us blamed for this, despite the fact that we drink less, take fewer drugs, commit less violent crime, and party less than previous generations did. Meanwhile, of course, in the great democratic exercises, in the elections, as our politics has become increasingly generationally divided for these socioeconomic reasons, I have seen the interests of my generation beaten resoundingly time and time again — in UK general elections, in US general elections, and on the Brexit vote. And, naturally, I have also witnessed the disastrous response to the COVID-19 pandemic, which — although I’ll admit is far less dangerous to young people — has sunk us into an even worse economic mire.

In exchange for this, I have received a smartphone which allows me to read more bad news stories at a faster rate, and the internet which has, at least, made it easier to try and find distractions in fantasyland from a reality that is distinctly not going that well.

And yes, I appreciate that former generations had their problems, as well, and that there have been recessions in the past, and that previous generations had to struggle to buy houses and get jobs when they were younger, I have absolutely no doubt about this. I’m sure that during the height of the Cold War, people were terrified of being incinerated by nuclear weapons. Although I would remind people that the nuclear weapons still haven’t gone away.

So I suppose I’d suggest that there may be some actual, concrete reasons that younger people are more depressed than ever which don’t just stem from increasing self-obsession, increasing awareness of the problem, or cultural trends, or the breakdown of earlier forms of community or family or religion that provided safety nets for people with these issues. But this is really another show.

Nevertheless, regardless of the cause, regardless of how valid you might feel it is, issues surrounding mental health are on the rise and are either becoming, or being recognised as, a serious societal problem. And, if nothing else, we look to technology to make our lives better — easier.

Plenty of the startups that I’ve looked at in this area are seeking to develop some kind of robotic, chatbot therapist. I did an episode about them recently which may or may not end up being released. We have apps for meditation. And we have prescription drugs. This is the only way that the systems of innovation, driven by profit and what’s easy to accomplish, know how to solve problems.

But if the reality simply is that the way we are living, or the way that we are governing ourselves, is making us miserable, is leading to all of this suffering — can tech really fulfil its promise of leading to a brighter future? When you’re promised Star Trek replicators that will make everything for you, but what you ultimately get is access to billions of YouTube videos, have you made a fair trade?

When it comes to climate change, as I will cover in much greater depth in other episodes, the positive futures that we have imagined increasingly depend on huge amounts of technological and technical innovation — even as the solutions that might actually help reduce emissions are taking the backseat to investments like Uber and WeWork.

In all of these visions that are set forth of the future, how much of our imagination are we really devoting to solving the grand challenges that actually face us as a species — and how much of our imagination is just dedicated to imagining that we can create an app that will destroy another industry, or hype up another company into multi-billion dollar IPO unicorn fantasyland? When you can destroy the business model of half of Softbank’s investments by asking what problem they actually solve, or what service they actually provide, you have to question whether we should have so much faith in our current set of innovations to solve everything.

We romanticise the founders of these companies as building new empires and often ignore or downplay the exploitative nature of the way that they do it.

And because there’s this feeling that we are so dependent on this growing sector to save us from all of our societal ills, we overlook when it does things that are stupid, or have terrible societal impacts.
Aside from maybe the European Union on occasion, no one seems willing to rein in the big tech companies, because they have convinced us that they hold the key to the future.

One exception here is China, which often seeks to replicate the same thing with companies that are under direct control of the government, but I don’t think that really counts, do you?

The reason I am so hard on the tech industry and the bubble it’s become is because, like a lot of people should feel, I am disappointed. It’s the disconnect between the lofty rhetoric of the tech companies and how their actions actually impact society. It’s the disconnect between Masoyoshi Son’s baffling, but fascinating, PowerPoint presentation and what he actually invested in. It’s the disconnect between the sci-fi, futuristic notion of tech in the sales brochure, and the technology that actually exists today. It’s the disconnect, in other words, between reality and the marketing material.
And the reason I feel disappointed is because, deep down, I still feel like technology is going to be a big part of the answer. Being a Luddite, or an anarcho-primitivist, or whatever, is not going to help.
We all know that technology has resulted in vast improvements to our life on this planet. It’s prevented us from dying at the age of 30. It’s allowed me to talk to all of you, through this show, hundreds of people I might never meet — when, in a pre-technical age, I’d know nothing and know even fewer people to tell it to. And I know that if there are going to be solutions to the growing set of problems that we face as a species — and, if we are going to find ways to make everyone’s lives better, and more worth living — technology and technological development is going to be a huge part of that.

Heck, even market forces will have their role to play, as I witness the incredible technological developments in solar power, in wind power, and in energy storage, which are finally driving fossil fuels out of the power generation industry in many places, even as other sources of greenhouse gases remain. I still believe all of these things to be practically self-evident.

But we can’t allow our shared belief in the crucial development of technology to blind us to the fact that the way technology progresses is not inevitable. The things we focus our time and energy on when we develop technology are not inevitable. The system of incentives that we set up to produce different kinds of innovation are not inevitable. They do not occur extrinsically to human decision-making processes. It is not external to the choices that people make, whether these people are programmers, tech CEOs, people who come up with and pitch startups, people who write about those companies, the venture capitalists that invest in their companies….

Evolution will happen. That much is inevitable. But there is not just one type of technological progress. The way we choose to guide it — the things we choose to emphasise and prioritise and optimise for, the behaviours we reward and the behaviours we try to suppress — these things are not inevitable. They are choices that we make as people. For far too long, we have been making the wrong choices.

Jacinda Ardern, the Prime Minister of New Zealand, recently made the observation that “economic growth while the quality of life stagnates is not progress.” To which I think we can add that “technological development that actively worsens quality of life for people is not progress.” And yet we maintain a great deal of faith that tech, or things that brand themselves as technological progress, is inevitable, profitable, and ultimately for the good of everyone.

So the question I want to ask — and the final question for this series — is whether that faith is misplaced. And, if so, how can we unblur Softbank’s Vision, and the vision of our society more generally? One of the hardest things to do right now is to imagine a utopian future. The tech utopias of the Victorian age — heck, even of Star Trek — have given way to an endless string of tech dystopias.

Too often, we struggle to imagine — with a straight face, and without an eye on the wallets of investors — how technology is going to develop that will genuinely make society better, safer, or generally more worth living in.

And this is why it matters: because of what we lost in the hype.

So how do we get that vision back?

www.physicalattraction.libsyn.com