OpenAI apparently asked the government for an investment backstop. This as concerns about the AI bubble are raised at the World Economic Forum.
When will the AI bubble pop? That’s a question on a lot of people’s minds as the parallels between the AI bubble and the .com bubble continue to be very apparent. On Tuesday, the stock markets took a pretty big beating as investors worried that the stocks were over valued. While the markets did recover, it marked at least the second time there was a scare in the market that the AI bubble had finally popped (it apparently did not).
Luckily, it did stir conversations that we are in a stock market bubble that could burst at any moment. In what has to be a case in particularly bad timing, while the markets experienced a huge drop, OpenAI’s CFO, Sarah Friar, was openly asking for investment guarantees. Basically, it was a financial backstop just in case things were to go wrong. The ask was seen by many as a pretty naked attempt to privatize the gains and socialize the losses – a major theme of the 2008-2009 financial crisis when banks asked to be bailed out by taxpayers because of really bad financial decisions on the banks part.
At any rate, the ask was so absurd, even the Trump administration didn’t even want to touch the idea with a ten foot barge pole. From CNBC:
Venture capitalist David Sacks, who is serving as President Donald Trump’s artificial intelligence and crypto czar, said Thursday that there will be “no federal bailout for AI.”
“The U.S. has at least 5 major frontier model companies. If one fails, others will take its place,” Sacks wrote in a post on X.
Sacks’ comments came after OpenAI CFO Sarah Friar said Wednesday that the startup wants to establish an ecosystem of private equity, banks and a federal “backstop” or “guarantee” that could help the company finance its infrastructure investments.
Following the comments, apparently OpenAI realized how ridiculous this made them look. They apparently suddenly found themselves walking back the comments, saying that when they asked for a government bailout, they weren’t asking for a government bailout, just, uh, an investing partner… or something…
She softened her stance later in a LinkedIn post and said OpenAI is not seeking a government backstop for its infrastructure commitments. She said her use of the word “backstop” clouded her point.
“As the full clip of my answer shows, I was making the point that American strength in technology will come from building real industrial capacity which requires the private sector and government playing their part,” Friar wrote.
See??? Totally convincing! Why would anyone take a request for a bailout be taken to mean that they want a bailout? That’s just silly. No one in their right mind would ever do that!
At any rate, the president of the World Economic Forum (WEF) is warning of three bubbles that are causing concern. From Reuters:
The world should watch out for three possible bubbles in financial markets, including artificial intelligence, the head of the World Economic Forum said on Wednesday, in comments that came amid sharp falls in global technology stocks.
Brokers and analysts say the falls are a cause for caution but not panic as markets have been touching record highs and some valuations are looking overblown.
“We could possibly see bubbles moving forward. One is a crypto bubble, second an AI bubble, and the third would be a debt bubble,” WEF president Borge Brende told reporters during a visit to Brazil’s financial hub, Sao Paolo.
Governments have not been so heavily indebted since 1945, he added.
Investors in the AI industry no doubt began to freak out that the the illusion of a boom just got pierced pretty badly, so they did what they do best: fire up the hype machine to try and paint over these concerns as best they can. From CNBC:
For Magnus Grimeland, founder of Singapore-based venture capital firm Antler, it’s clear the market is not overheating. “I definitely don’t think we’re in a bubble,” he told CNBC’s “Beyond the Valley” podcast, listing several reasons.
The speed at which AI is being adopted by businesses is notable compared to other tech shifts, Grimeland said, such as the move from physical servers to cloud computing, which he said took a decade. Added to this, AI is “top of the agenda” for leaders today, he said, whether they’re running a healthcare provider in India or a U.S. Fortune 500 company.
“There’s a willingness to invest into using that technology … and that’s happened immediately,” Grimeland said.
He described the rapid shift to AI as being substantially different from the dotcom bubble of the late 1990s and early 2000s, when unprofitable internet startups eventually collapsed and the tech-heavy Nasdaq
lost almost 80% of its value between March 2000 and October 2002.
“What makes this a little bit different from a bubble and makes it very different from dotcom is that there’s really real revenues behind a lot of this growth,” Grimeland said.
Yeah, let’s rely on the “it’s different this time” strategy. After all, it’s not like denial was ever a thing in the .com bubble or anything like that. I mean, that alone tells you that it’s totally not a bubble, right? Right???
The arguments that we’re not in a bubble are ultimately unconvincing. People were rapidly adopting the internet as a technology, there were products being sold online, yet the .com bubble popped anyway. The reality is that companies are having a hard time selling their AI products thanks, in part, to the technology simply not living up to the over-promised hype. A huge chunk of the money being invested has been flowing from one AI company to another to try and pump up the numbers. What’s more, the AI industry is also largely fuelled by hype rather than actual booming sales to back up such wild claims.
Obviously, it isn’t just me saying these things. Recently, an FAQ was published on The Ringer making many similar observations:
How bad is the mess? OpenAI’s ChatGPT, by far the most successful generative AI product to date, requires so much expensive computing power to run that it loses money almost every time you use it. (For the first half of 2025, OpenAI took in $4.3 billion in income and still reported a net loss of $13.5 billion.) A recent Deutsche Bank report says that the AI boom is “unsustainable.” Bain & Co.’s annual global technology report says the AI industry will need $2 trillion in annual revenue by 2030 to continue at its current rate. (That number will be, um, hard to come by.) A Bank of England report warns that the market could experience a “sharp correction” due to the overvaluation of AI companies.
Companies with no clear plan for making money from AI have poured cash into the construction of data centers so massive that they’re straining the electrical grid and driving up the price of electricity for the rest of us. They’ve poured cash into what’s called “compute”—processing power, basically, used to train AI models and run search queries—at a rate they can’t keep up unless the entire world economy reorganizes itself around their products by, like, Tuesday. All that tech has a shelf life, just as your iPhone does; the industry will have to keep spending ruinous sums every few years to replace it. And to keep the ball in the air, the companies have made a number of increasingly bizarre, shady, and circular deals with each other: Nvidia, for instance, has invested $100 billion in OpenAI, which OpenAI must then turn around and spend on Nvidia products. If I give your lemonade stand $10 so you can buy my $10 lemons, we can’t tell our investors (your mom) that we’ve boosted the lemonade economy by $20. But that’s the kind of reasoning that fuels more and more of Silicon Valley’s grandiose claims about itself.
The thing is, though, the price of a stock is not determined by the soundness of the company. It’s determined by the willingness of investors to pay a certain price. And there are ways to convince investors to pay high prices for shares even when it means driving through an endless series of flashing red lights. If you can, say, leverage three decades of mythmaking about the transformative impact of the tech industry by convincing a compliant press to spread the word that your new product is poised to make every sci-fi dream a reality—cure cancer, bring humanity to other planets, create flying cars—then the sheer size of the vision you’re peddling might convince investors to buy in, even if doing so seems reckless.
Tech investors, especially the superrich ones, are always in the market for stories scaled to their egos, and their egos are getting more messianic all the time. And in the era of crypto and Robinhood, regular idiots like us increasingly see the stock market as a casino. Pushing some chips onto the new, hot, famous, sexy thing seems way more exciting than researching the fundamentals of some well-run company no one’s talking about. Plus, everyone else is doing it. OpenAI is currently targeting a $1 trillion IPO. The basis of that number is not—like, extremely not—the solid state of the business. It’s the story the company is telling.
As you can see, I’m not the only one talking about how some members of the press are doing people no favours on this front. I’ve long argued that the media outlets are simply taking the tall tales told by the AI industry and running those stories as fact without really questioning it. Here, The Ringer is basically saying the same things.
When this whole thing comes crashing down, it’s going to hurt… a lot. The longer this bubble continues to inflate, the harder the crash is going to be. As a result, the sooner this bubble bursts, the better.
Drew Wilson on Mastodon, Twitter and Facebook.
Discover more from Freezenet.ca
Subscribe to get the latest posts sent to your email.
