Where are All the AI IPOs?
- Karl Johansson
- för 3 dagar sedan
- 6 min läsning
Why has OpenAI and Anthropic avoided going public when they are some of the most hyped firms on the planet who also need to raise a lot of money to fund their capital expenditures?
In spite of all the ways AI is a product of the computer age it is at its core an industrial project. We tend to think of AI labs as the quintessential Silicon Valley scale-ups: offices in Palo Alto where college drop-outs and brainy computer scientists create the future. But in reality the models which underpin the AI eco-system are dependent on huge amounts of chips which are housed in buildings complete with HVAC, electricity, pipes and other accoutrements more familiar to mid-century industry than the new millennium’s code workshops. The big issue in AI and tech more generally is a familiar one to students of industrial history: how to afford big machines required to get production off the ground.
The solution Western capitalism found was to create companies with publicly traded shares. Sell a share of the company to others and use the money gained to invest in the factory, creating a partnership of investors and industrialists who both gain from the construction of new industry. As the computing and the internet took off, a new type of company was created which thrived on making the economy more efficient through connection and which competed not through what it produced but through what it facilitated. That type of company had less use of going public as the start-up costs were both lower, and crucially, the path to profitability was shorter and less contingent on pesky things like supply chains and finding a workforce both capable and big enough to man the factory floor. To get a factory going you need to get from 0-80 before it can sustain itself, whereas to get a Silicon Valley tech firm going you only need to get from 0-30. Coupled with the aforementioned lower costs to start generally the financing options are far wider than for say a car factory.
But now that the Valley needs AI factories it has not turned to the public markets to finance it, but has instead continued using venture capital. Why?
The first and most benign reason is good old fashioned path dependency. The AI whizz kids, Sand Hill Road venture capitalists, ex-McKinsey MBA’s, and the tech tycoons are unused to dealing with IPO’s and default to working the way they always have, which means relying on private companies and venture capital. To the veterans of the post-2008 tech boom an IPO is what you do to cash out your chips; it is not a valid move in the game of tech investing.
For the past two decades that strategy has served the tech industry well. Few if any tech companies have required the amounts of capital old industrial giants used to, and when they did need a lot of money interest rates were generally low. So during the 2010’s access to capital was never an issue, as each successful venture capital deal created more money for the VCs to use in future investments and more interest from institutional investors for shovelling money to unlisted tech firms. Indeed, the whole point of venture capital is to make smallish bets on many low probability high impact start-ups, and therefore the risk-reward profile only works when the bets are small.
The financial infrastructure which has propelled the growth of the Valley’s software magnates was never built to finance billion dollar capital expenditures. The relationships techies have are with the wrong type of financiers, and they have not done deals like this themselves, nor are they likely to know anyone who has. Is it really strange that they haven’t thought to IPO?
A second, slightly more cynical explanation is simple greed. If you take the AI CEOs and thought leaders by their word, they sincerely believe that they are in charge of the most impactful, important, and impressive technology in history. That confers wealth and power, which you might not want to be overly generous with. If you accept the premise that essentially all coding and at least half of all white collar work is going to be automated in one to two years (it is always one to two years) then it makes even more sense to jealously guarding a potential source of incredible wealth.
AI believers disagree on whether advanced/superintelligent/general AI (I forget what the fashionable term is this week) will be great or terrible, but they all agree that it will be a big deal with tremendous social, political, and economic ramifications. Navigating that upheaval will be easier and more comfortable if you and your family’s finances are secure for at least a couple of generations, so diluting your equity stake in history’s biggest deal could be construed as a rookie error.
Returning readers will know that I find all the hysterical and/or pompous philophisising about the post-singularity future to be a complete waste of time, as there are no reasons to believe that these promises will come true. The discussions about AI are often warped because believers pass off the burden of proof to sceptics rather than taking it on themselves; insisting that experts agree that a godlike machine is right around the corner, and trying to make you the unreasonable one for not simply accepting their claim without proof or evidence.
Still, it would not be very surprising if the AI evangelists made a conscious effort to avoid IPO’s in preparation for the future they hope to serve as the midwives for. It is cynical, but far less cynical than the alternative: that they don’t think the grift of raising venture capital would work if people could see just how rotten the finances truly are.
When you really get down to it, is there any good reason for OpenAI and Anthropic not to have gone public? They keep getting billions of dollars every few months with the media reporting valuations ratcheting ever upwards. Why not raise a lot at once through public markets, it is not as though they cannot sell more shares if they want or need more money after an IPO? Why waste the time and effort of courting private capital all the time when they are literally the most talked about businesses in the world, and have been for a couple of years now?
Why are they not just listing on the Nasdaq? It is not because they are too small for public markets, having raised tens of billions of dollars they can deal with all the reporting requirements. It is not because there is insufficient interest. It is not because tech valuations are low at the moment. So what is it?
My guess is that they are loss-making enterprises with fundamentally bad unit economics which would not be able to keep the lights on if they were not subsidised by venture capitalists. I could be wrong; indeed I hope I’ll be proven wrong as that would be a rebuke to my cynicism regarding the American technology sector. I will of course acknowledge my error if and when OpenAI and Anthropic are able to show that developing AI models is good business.
Going public is not the be all end all of capitalism, it is not necessary to building a great business. But it undoubtedly seems like the best path forward for the AI labs. Every investment bank in America, perhaps in the world, would be overjoyed to get to be the ones to take OpenAI public. Not only would it be a juicy deal in terms of profit, but it would make the careers of those who worked on the deal. I guarantee that Sam Altman has been approached by a baker’s dozen banks hoping to get to be the ones who take OpenAI public.
Why hasn’t he taken them up on the offer? My guess is that unlike the industrial capitalism of old, OpenAI’s capital expenditure is not productive. That is the only reason I can come up with why the AI labs won’t use the answer we have already found to the question of how to afford building industry. The server farms he wants to build are only ever going to manufacture hype, and while that is most certainly a mass market product, it is not one which durable businesses are built on.
If you liked this post you can read a previous post about Ukraine & Iran here or the rest of my writings here. I also have a section for longer reads I call essays here, I particularly recommend my series called The Bird & The Technoking exploring Elon Musk's takeover of Twitter, and its political and cultural implications. It'd mean a lot to me if you recommended the blog to a friend or coworker. Come back next Monday for a new post!
I've always been interested in politics, economics, and the interplay between. The blog is a place for me to explore different ideas and concepts relating to economics or politics, be that national or international. The goal for the blog is to make you think; to provide new perspectives.
Written by Karl Johansson
Kommentarer