1) In last Monday's e-mail, I said that the current AI craze reminded me of the Internet bubble and concluded:
There's little doubt in my mind that we're in a similar AI bubble right now – but I'm much less certain how much further it will inflate and when exactly it will burst.
Here's one more similarity that makes me think we're near the peak...
CNBC's Jim Cramer recently visited a construction site in Idaho for a Micron Technology (MU) semiconductor plant. He complained that "there's an incredibly jarring gulf between stock prices and reality" and that "Micron's stock is radically undervalued."
This is significant because, for decades, Cramer has been a great contra-indicator.
It reminds me of his infamous "Winners of the New World" speech on February 29, 2000, just 10 days before the Nasdaq Composite Index peaked.
In it, he named 10 Internet-related stocks and said:
We try to own every one of them. Every single one. And if I had my druthers, I wouldn't own any other stocks in the year 2000. Because these are the only ones worth owning right now in this extremely difficult, extremely narrow stock market. They are the only ones that are going higher consistently in good days and bad. I love every one of them, just as I loathe the rest of the stock universe.
Within 15 months of his speech, the basket of stocks lost more than 80% of its value. By 2009, all 10 companies were either bankrupt, delisted, or acquired for a tiny fraction of their peak bubble valuations.
I think history is likely to repeat itself with the AI bubble.
2) I've been warning my readers about ChatGPT operator OpenAI for years, calling it "a cash-burning furnace" (archive here). And things continue to go from bad to worse for the company...
As reported in the Wall Street Journal, its second-quarter revenue was up only 18% from the first quarter. Meanwhile, revenue more than doubled for rival Anthropic.
At the beginning of this year, OpenAI had twice as much annual recurring revenue as Anthropic. But it has now fallen far behind, as this chart from TMT Breakout shows:
And the cash-flow trends are even more starkly divergent, as you can see in this chart created by my friend James Emanuel:
These trends are why I think Anthropic will have a successful IPO, perhaps as soon as October, whereas I doubt OpenAI will ever go public. Instead, I think it's likely to implode spectacularly.
If I'm right, the consequences will go far beyond one company, as blogger Ed Zitron writes here: What Happens If OpenAI Dies? He summarizes why he's so bearish on the company:
OpenAI has no economies of scale, it's horribly-unprofitable, and does not have a stable business. This naturally means that it has to continually raise capital, except raising further capital is going to be difficult, based on the sheer amounts it needs, the dwindling funds available for it to raise, its already-inflated valuation, and the fact that it's way behind a competitor facing exactly the same problems.
OpenAI has promised the impossible, and built a company that only makes sense if you're willing to ignore the worst economics in the history of capitalism. Its future is dependent on raising over a hundred billion dollars a year in one of the worst funding climates in history. Its revenues are slowing, its competitor (and there's really only one) has outpaced it (all while slowing itself), and its CEO is one of the single-worst spokespeople in history...
The collapse of OpenAI would likely be a result of the walls closing in around its ruinous obligations and economics, with counterparties left short-changed and deals broken as things begin to unravel.
How might the unraveling play out? Zitron outlines a number of scenarios, but I think this is the most likely one:
While I imagine some rescue package is pulled together, OpenAI could simply be allowed to run out of money, short-changing nearly a trillion dollars' worth of compute contracts, killing CoreWeave, Cerebras, and anyone else reliant on its income. Its customers would be given API keys that flow to Microsoft AI Foundry, Amazon Bedrock and Google Vertex, and be told that there would be little or no further development or training of OpenAI's models.
This situation, while obviously destructive for the entire industry, would give everybody a scapegoat. Who made all the promises? Sam Altman. Who ran a shitty company into the ground? Sam Altman. Who misled everyone into believing that there'd be infinite demand for compute? Sam Altman. Stories will leak that OpenAI was "not consistently candid" with its financial condition with partners, allowing everybody to reframe a trillion-plus dollars in waste as the result of one egregious con artist.
3) If Zitron and I are right that OpenAI is going to be a train wreck, the first stock to go to zero will likely be data-center builder CoreWeave (CRWV).
My friend Porter Stansberry described its awful economics in a blog post on May 13:
CoreWeave is the cleanest case study of how financing has built this bubble. It has raised $28 billion in equity and debt in the last 12 months.
The company closed an $8.5 billion delayed-draw term loan in March. That was the first investment-grade financing in history secured by GPU hardware and customer contracts, rated A3 by Moody's. Before March 2026, no one had ever managed to convince Moody's to give an investment-grade rating to a loan backed only by GPU chips. Historically, GPUs were considered too volatile, too short-lived, and too easily made obsolete to support an investment-grade rating. And Moody's gave this line of credit an A3 rating – that's three notches into investment grade. That's the kind utilities and railroads get.
The bond matures in March 2032. That's a six-year maturity against assets (Hopper-generation GPUs) that Nvidia (NVDA) is already several development cycles past. This is a six-year loan against an asset with a two-to-four-year life span. The cash flows depend on payments from a company (OpenAI) that, by its own backers' projections, is going to lose $35 billion in 2027.
In an X post on August 1, he summarized the company's financials:
CoreWeave: $5.13 billion of 2025 revenue, $14.9 billion of capital expenditure, negative $7.25 billion of free cash flow, net debt at 8.1 times EBITDA, term loans at 11% to 15%, a weighted-average short-term borrowing rate of 12.3%, and a $1 billion private placement in April 2026 at 9.75%.
For a more in-depth bear case, see this Substack post by Alexandra Damsker: Nobody Tell CoreWeave It's Not a Business. Excerpt:
On July 29, credit default swaps on CoreWeave debt priced a 50% five-year default probability, a premium level normally associated with deeply distressed borrowers. Not a company that was rocketing to revenue highs. Credit markets are pricing CoreWeave like a company on the brink, while stock markets are looking at it like a rocket ship. That's not a great sign...
Watch the credit market, not the stock price, for the real signal on CoreWeave. Equity markets love revenue numbers, and focus on those happy, curated earnings calls. Credit markets focus on the probability of getting paid back.
When those two markets disagree this sharply, I always go with credit, because it actually cares about diligence and holds more professional players. The CDS spread here is definitely the earlier and more honest indicator, and it has been a flashing red "check engine" light since before this week's rally. A 50% five-year default probability sitting underneath a stock that just popped 14% is not a bet the stock market figured out. It's a contradiction that one side is clearly misinformed about, and between the credit and equity markets, the equity markets are operating optimistically blind and with more downside.
Let's take a look at CoreWeave's historical financials, which are quite limited because it's less than 10 years old. (It was originally called Atlantic Crypto Corporation – talk about a red flag! – and changed its name to CoreWeave in December 2019.)
Revenue growth has been strong, but the company barely breaks even on an operating-profit basis. And it has had negative net income every quarter thanks to high, rising debt payments:
Analysts don't expect the company to turn profitable anytime soon. Consensus earnings estimates are negative $5.20 per share this year, narrowing slightly to negative $4.24 next year.
CoreWeave has generated a bit of operating cash flow, but this is far exceeded by its capital expenditures ("capex"). As a result, free cash flow ("FCF") is accelerating to the downside:
CoreWeave funds its huge cash burn by borrowing more and more money, resulting in rapidly rising net debt:
These are among the ugliest financials I've ever seen...
Net debt of $46 billion is nearly equal to CoreWeave's market cap of $48.5 billion, giving the company an enterprise value of $94.5 billion. That's ridiculous.
CoreWeave investors are betting that the AI boom will expand even further, allowing the company to turn profitable and pay off its enormous debt load. Not likely...
I think what they're really betting on is that someone will buy their stock at a higher price at some point in the very near future – the very definition of speculation.
So today, I'm adding CoreWeave to my "Stinky Six" list of stocks to avoid and renaming it the "Stinky Seven."
Best regards,
Whitney
P.S. I welcome your feedback – send me an e-mail by clicking here.
P.P.S. Longtime readers know I'm a big tennis fan. I play on a number of U.S. Tennis Association teams and try to go to a few tournaments every year. But I probably won't go as often as in past years, as the tickets have become shockingly expensive. For example, a grounds pass for the U.S. Open on Sunday is currently around $300 on the reseller sites.
But here's a tip: It's Fan Week now through Saturday, during which admission for most events is free! You can click here to register for a pass to see 256 players compete in the qualifying tournament for 32 slots in the main draw.
Today, tomorrow, and Wednesday is the mixed doubles tournament, which has attracted many top players – including Alcaraz, Djokovic, Zverev, Tiafoe, Fritz, Serena Williams, Pegula, Sabalenka, and more. You can see the draw here and schedule here.
I'll be there tomorrow at 10 a.m. to cheer on my favorite team: Elina Svitolina, whose foundation I support, and her husband Gael Monfils, one of the most popular players on the tour, who's retiring this year after a 22-year career.
This is the first time they'll be playing together in a real tournament. Both of them are in the singles draw – she's ranked No. 9 in the world, and he received a wild card. They'll be facing off tomorrow against Alexandra Eala and Felix Auger-Aliassime.
I had the pleasure of spending the day with Monfils yesterday. He just started playing golf a month ago and has gotten hooked, so I arranged a round for him at the beautiful Quaker Ridge Golf Club just outside New York City. (Thank you to our host, Andrew K.)
Monfils is looking to play again later this week, so if you're a member of a club anywhere near the city – or have a connection at one – and can get him in, please e-mail me!
