Last week, I laid out the case for why I think the AI boom has turned into a bubble and is starting to burst – just like the Internet bubble did in March 2000.
But this will be nothing more than a hiccup for the best companies that generate gobs of cash, like Meta Platforms (META), Amazon (AMZN), Alphabet (GOOGL), and Microsoft (MSFT) – all of which I analyzed on Monday.
At the same time, it's very bad news for junk companies with access to nearly unlimited quantities of cheap debt, like CoreWeave (CRWV). Just as I predicted, the stock is imploding, which I discussed on October 15.
More evidence that supports my view on the AI bubble comes straight from Meta's recent 14C filing. When a cash-gushing company like Meta is worried about raising money, imagine what it must be like for just about everyone else:
Recent significant capital raises at other notable companies, and others soon anticipated to take place, to fund the same AI growth – for example, Alphabet, SpaceX, Anthropic, and OpenAI – have resulted in an increasingly challenging environment for raising capital.
This has investors spooked, as this story on the front page of today's Wall Street Journal notes: Meta's Hunt for AI Capital Strikes a Raw Nerve on Wall Street. Excerpt:
The vast amount of capital needed to fund the artificial-intelligence build-out is forcing tech giants to hunt for ever-larger sums of cash – and Wall Street is raising a bright red flag.
Signals from Meta Platforms that it could soon go on a cash-raising spree helped trigger another massive rotation trade on Thursday, with investors dumping stocks connected to the AI boom and piling into almost everything else, from U.S. Treasurys to the beaten-down shares of home builders and software companies...
The big moves came after Meta announced plans Wednesday to create a new class of nonvoting stock to raise the capital it needs for developing and serving AI models. Initial excitement over the company's record third-quarter revenue quickly gave way to mounting anxiety over how the Facebook parent and its peers will secure enough capital to fund the AI race.
"The growing concern with AI isn't about technology anymore," said Michael Antonelli, a market strategist for Baird Private Wealth Management. "It's 'how can we all fund this at once?'"
The answer is simple: It can't all be funded at once.
That's evident through the credit-default swap ("CDS") spread, a measure of how likely investors think a company is to default on its debt.
In the chart below, made by a friend of mine using Bloomberg data, you can see that CDS spreads are blowing out for CoreWeave, Oracle (ORCL), and SpaceX (SPCX), but not for Amazon and Meta:
Note: CoreWeave's data in basis points ("bp") is on the right-hand scale, and its spread is more than three times Oracle's.
I'm concerned about three types of circularity: circular financing, circular valuation pumping, and circular demand.
This famous chart from a January Bloomberg article, A Guide to the Circular Deals Underpinning the AI Boom, shows the degree to which the major players in the AI space are investing in and buying from each other:
In doing so, these companies are creating a valuation bubble. One dollar spent by an end user is turned into many times more in valuation.
This X post by Wing Venture Capital's Zach DeWitt gives a good example of a how each dollar of a law firm's spending on Harvey, a privately held AI service, turns into $100 of asset value:
This is indeed "one of the greatest flywheels in capitalism" – but think about what happens if end-user demand slows and the flywheel spins in reverse...
As for circular demand, journalist Derek Thompson does a good job of outlining the issue in this Substack post: The AI Story Wall Street Doesn't Want to Hear. Excerpt:
But as tech stocks wobble, some AI skeptics are scrutinizing another number that is frustratingly hard to pin down: How much external revenue is the AI ecosystem actually generating?
This is a surprisingly hard question to answer. The supply side of the AI economy is fairly legible. Nvidia reports its revenue every quarter. Microsoft, Amazon, Meta, and Alphabet report their capital spending. But the demand side is more opaque – a dark archipelago of private AI model makers, such as OpenAI and Anthropic, that don't yet report quarterly revenue figures. AI usage is often bundled into products like Microsoft 365 and Salesforce, which means even public-company earnings do not cleanly indicate how much of their revenue came from AI.
The murkiness of AI revenue is starting to worry analysts. "All I care about is, are the revenues coming into the system greater than the costs of building the whole value chain," the tech and media analyst Craig Moffett told Ben Thompson. When chipmakers sell to clouds, who sell to labs, who buy chips, who sell to clouds, that's just money changing hands between layers of the value chain, he said. What matters more is the money coming in from the outside – from individuals, businesses, and governments. "The market is panicking right now," Moffett said, "because the market is saying, 'We don't know.' It's not obvious."
On one hand, both trailing and annualized run-rate AI revenue is soaring, as this chart in Thompson's article shows:
However, AI spending is highly concentrated among the top 1% of companies. They spend a median of nearly $10,000 per employee per month, accounting for 80% of total spending:
Meanwhile, the median user only spends $15.70 per employee per month. I'm a good example of this: I pay $20 per month for Anthropic's Claude and am very happy with the basic service. I don't have any reason to spend more.
So, who are these mega-users? Nearly all of them are venture-backed companies, as shown in this chart by Torsten Slok, the chief economist at Apollo:
This is very worrisome, as Thompson points out:
Why does this matter? Well, AI demand is supposed to account for revenue generated outside of the AI capital cycle. But if most AI spending comes from companies that rely on venture capital firms that rely on AI valuations, this seems like another example of the AI capital cycle paying itself. If the AI buildout wobbles, and venture returns suffer, it could force the biggest AI spenders to pull back at the very moment that the AI project would need their revenue the most.
And indeed, AI spending by mega-users appears to have peaked and is now leveling out:
But interestingly, their usage is still going up. That's because prices are coming down:
As Thompson explains:
The first half of 2026 was the brief era of tokenmaxxing, as companies ran up wild AI budgets without proof of [return on investment]. But in the last few months, companies have been more discerning about using expensive models for advanced tasks and cheap models for simple tasks, as Moffett told Thompson:
"If you're doing something at the bleeding edge of research – like I'm doing protein folding for some next drug compound – you may absolutely need the best model. But if you're saying, for my enterprise I want to transform customer service, I don't need the best model, I'm going to use a cheaper model that's two or three generations old, and I think there's going to be a lot of price deflation."
As a result, the effective price that businesses pay per million tokens fell 41 percent between March and August.
The mistake investors are making is a classic one – they're projecting the immediate past indefinitely into the future. Thompson agrees:
In 2020 and 2021, investors looked at Peloton's (PTON) and Zoom's (ZM) explosive growth, projected their revenue curves into the future, and concluded that these companies would take over fitness and work. Instead, what the pandemic really did was concentrate several years of demand into a few quarters. The earnings that followed disappointed, and their stocks plummeted.
Financial-intelligence firm Vital Knowledge reaches a similar conclusion, which my friend Doug Kass shared in his Daily Diary (for subscribers):
Amid all the noise around AI, the underlying problem for the industry is one that's been present for a while: AI growth is dependent on capex (this is one of the most capital-intensive industries ever) and since capex is dramatically outpacing operating cash flow, companies are reliant on the largess of equity and debt investors to underwrite all the spending, but lately those markets are becoming increasingly less accommodating.
Because the issuance landscape for debt and equity has deteriorated materially, this will inevitably lead to a reduction in aggregate spending (the pool of capital willing to pay for everything is smaller than before, at least for now).
It used to be that investors could focus myopically on the big headline AI numbers (user growth, [annual rate of return], backlog, etc.), but that no longer suffices as these top line metrics are impossible to sustain/realize without funding, and that funding is now a lot more precarious...
The argument can be made that this funding issue will soon pass because underlying demand is so strong, but it's important to note that a lot of the demand is a function of that funding (i.e. loss-making companies are fueling a big chunk of token consumption).
As it continues, the AI sector now depends on two main entities:
It can't be overstated how important Anthropic and OpenAI are – they remain the suns around which the whole AI world rotates and if these companies can't come public on the terms anticipated by bulls (in terms of timing and valuations), there will be negative fallout for everyone due to 1) less money available to spend on compute and 2) a lower benchmark for sector valuations.
Compute is fungible, and the capacity being built for Anthropic/OpenAI could be repurposed over time if they don't fill it, but that process won't be quick or painless.
In summary, I'm bullish on AI technology. I think it will be as transformative as the Internet, and we'll all massively benefit from its implementation.
At the same time, the largest bubble in history has formed around this technology. Now, as capital markets and mega-users pull back, it's starting to burst.
This will result in a lot of pain in the sector – especially for speculators. And it will create some huge losers out of the junk companies that continue to pile on debt.
But the strongest AI companies won't just survive... They'll go on to dominate the market for years. Like I said, established, cash-gushing businesses like Meta, Amazon, Alphabet, and Microsoft could end up as long-term winners.
My friend Jeff Brown, publisher of Brownstone Research, has identified another of these AI winners... He's calling it the "AI Super Chip." And I'll be joining him to introduce the stock during a special presentation on Wednesday, October 28.
You don't want to miss this event. To reserve your spot, click here.
Best regards,
Whitney
P.S. I welcome your feedback – send me an e-mail by clicking here.