For years now, experts and politicians have been predicting that the apparent AI bubble would soon burst.
Companies have invested hundreds of billions of dollars in fancy new data centers and absurdly well-paid research teams in hopes of building powerful and wildly profitable AI models.
This is despite the fact that even the most innovative AI companies still have modest revenues. OpenAI will make only $20 billion in 2025, less than what Ross Department Store makes selling clothes and about the same as Frito-Lay makes selling potato chips.
Given those income realities, the current absurd level of investment seems unsustainable. But if OpenAI’s massive new funding round is any indication, the AI bubble isn’t going to burst. At least not yet.
Billions and billions
In late February, OpenAI announced that it had raised $110 billion to continue building its world-leading frontier models.
The deal values OpenAI at between $730 and $840 billion. And perhaps most importantly, the round was led by large companies and institutional investors, giving that valuation some real credibility.
Amazon reportedly contributed $50 billion, NVIDIA contributed $30 billion, and SoftBank matched NVIDIA’s contribution. In turn, OpenAI agreed to use Amazon’s cloud infrastructure for some of its products and secured access to more advanced AI chips from NVIDIA.
OpenAI said the deal would allow it to scale “AI for all” while funneling more capital into its nonprofit OpenAI Foundation.
The deal is especially notable because OpenAI remains a private company. In the old days (which in Silicon Valley means “a decade ago”), companies had to go public to raise this type of funding.
Now, companies like OpenAI and Elon Musk’s newly merged xAI/SpaceX can drive $1 trillion valuations without needing to involve pesky public market investors or accept the scrutiny that comes with an IPO.
Of course, OpenAI could always go public later (as many expect it will eventually happen) and likely raise hundreds of billions more.
What bubble?
OpenAI’s huge cash influx is also notable because some industry experts thought it might never happen.
NVIDIA reportedly toyed with the idea of investing $100 billion in OpenAI on its own, before it appeared to back out.
That pushback, along with OpenAI’s very strange governance structure, made some analysts nervous.
Stansberry Research’s Whitney Tilson wrote about a month before OpenAI’s February round was announced that “…’seeking’ $100 billion is very different than ‘receiving’ $100 billion,” and predicted: “When (not if) OpenAI fails to raise the money it seeks, that will be a key indicator that the bubble is bursting, and it will be time to exit.”
Similarly, analyst and New York University professor Gary Marcus told Goldman Sachs that “current AI companies aren’t generating enough profits to justify their lofty valuations. People are starting to get that message. And if enough people get that message on the same day, it’s going to start to look like a run on the bank.”
They shouldn’t have worried. OpenAI appears to have had no problem completing one of the largest private funding rounds in history.
And more investors, including tons of retail investors who would love something to invest in besides GameStop and Tesla, are almost certainly waiting in the wings.
All of this suggests that fears of a funding failure (and the subsequent bursting of the AI bubble) were probably overblown. OpenAI now has enough cash to continue burning billions for at least months.
The fact that OpenAI continues to pour money into building AI infrastructure also means that larger (and much richer) competitors like Google and Meta will be practically forced to do the same, or risk uncomfortable questions from their own investors.
Therefore, the AI money gravity train will continue to move forward, as the AI funding flywheel shows no signs of slowing down.
where it ends
I will take a moment to point out that nothing contained here is investment advice and you should not trade based on anything contained in this article.
This is especially true because despite the positive outlook offered by the gigantic rise of OpenAI, the fundamental business challenge facing generative AI remains unchanged.
Once again, OpenAI’s revenue is negligible. Yes, it is growing. But so is McDonald’s.
As multiple analysts have noted, investors are happy to overlook that and throw wads of money at OpenAI and the like because they hope generative AI will take over the job market, render humans obsolete, and usher in an era of transformative leisure in which we all swim in lakes while benignly intelligent supercomputers run the economy for us.
However, the moment that future shows signs of not appearing, investors like Amazon could decide to back off and put their money into more practical things, like selling toasters and ruining the legacy movie theater.
That moment didn’t happen this month. But it could still be right around the corner.
“AI is certainly in a financial bubble,” Marcus told Goldman Sachs. “Just as no one could predict the exact moment when the Dutch tulip mania would end, it is impossible to know when this bubble will burst, but it will burst.”