1) In last Tuesday's e-mail, I covered AI giant Anthropic's S-1 filing for its IPO, which will likely be near the end of the month.
The company reported incredible numbers. Revenue grew 1,368% year over year ("YOY") in the second quarter and 144% from the previous quarter.
But it didn't provide third-quarter financials. As I followed up on in Wednesday's e-mail, this led skeptics to question whether the numbers might not be so rosy.
Sure enough, the company filed an amended S-1...
It disclosed that revenue growth slowed dramatically – from 144% in the second quarter to 47% in the third quarter. Total revenue increased by $6.82 billion in the second quarter, slowing to a $5.4 billion increase for the third quarter.
Part of the reason for the slowdown was roughly $600 million of price reductions from new AI model versions and contractual discounts – not a good sign.
As for profitability, Anthropic swung to a $567 million operating loss, compared with a $974 million operating profit in the second quarter.
Of greatest concern are the company's staggering compute commitments...
In its original filing, it reported nearly $590 billion. But in the amended S-1, that figure is up to $895.4 billion. Nearly $70 billion of this is due over the next 12 months and roughly $448 billion before the end of 2030.
The total is equal to roughly 13 times annualized third-quarter revenue. No wonder the company is rushing to go public, seeking to raise an unprecedented $100 billion.
It had $146.6 billion in cash and readily accessible investments as of the end of the third quarter. But Anthropic is still going to have to raise or generate two-thirds of a trillion dollars to fulfill its contractually binding compute commitments.
That's going to be even more difficult if Anthropic's debt is junk-rated...
According to this TechCrunch article, investment giant Blackstone (BX) initially said the company would receive an investment-grade rating. But now it expects Anthropic to be rated BB, which is "high" junk.
I'm not surprised by this report in the Wall Street Journal yesterday (before the amended S-1): Anthropic Investors Push for Cautious Valuation Closer to $1.5 Trillion. Excerpt:
The rout in technology stocks last week has big investors who plan to participate in Anthropic's IPO coalescing around a lower valuation for the artificial-intelligence giant.
In the past several days, institutional investors told the Claude maker and its advisers that the company's valuation in its planned initial public offering later this month should be around $1.5 trillion to $1.6 trillion, according to people familiar with the matter, well below figures closer to $2 trillion that had been bandied about in recent months.
The investors cite a list of concerning factors including choppy markets and growing concerns about overexposure to AI bets. Political backlash to data centers and uncertainty around how soon hyperscalers such as Amazon.com and Meta Platforms could show AI-related profits are also weighing on investors' minds.
For more insight, see this Prof G Markets interview Ed Elson did with John Foley, head of the Lex column at the Financial Times: Anthropic's Q3 Results Are In – The Costs Are Exploding.
In summary, Anthropic's latest filing reinforces my view that the AI bubble is in the process of bursting.
If the best company in the sector is seeing slowing revenue growth, price cuts, profits turning to losses, exploding compute commitments, junk-rated debt, and valuation haircuts... imagine how bad things are for OpenAI, a far inferior company, in my opinion.
And, to be clear, these two companies are the pillars of the entire AI sector.
I question whether OpenAI will ever go public and have compared it with WeWork, as I mentioned in yesterday's e-mail. WeWork's IPO attempt ended in spectacular failure five years ago after its private valuation plummeted from $47 billion to less than $10 billion in a matter of weeks. It filed for bankruptcy two years later.
(Fun fact: WeWork founder Adam Neumann and OpenAI founder Sam Altman were both about 30 years old when they established their companies. I see parallels in their inexperience and youthful arrogance.)
As I've been warning, when the AI bubble bursts, there will be plenty of losers. But my friend Jeff Brown, founder of Brownstone Research, has identified a potential winner...
This company's AI "super chip" is 50 times faster than Nvidia's (NVDA). And it could open the opportunity of the decade for investors who get in early – before a pivotal announcement on Thursday, November 5.
Jeff and I went on camera last week to share all the important details. But if you missed our presentation, don't worry – you can watch a replay right here.
2) Another warning flag for the AI sector appeared on Sunday...
An X user asked Tesla (TSLA) and SpaceX (SPCX) CEO Elon Musk, "Any chance for a Grok Bot reset, @elonmusk? Lots of people say their usage ran out really fast this week." To which Musk replied:
Last weekend, some Grok Bot users found their weekly usage limits maxed out and thought it might be a metering bug. But Musk said it was a deliberate tightening due to a cost-revenue imbalance... In other words, SpaceX was losing a ton of money.
Musk's comment was jarring because he'd posted this three days earlier:
And during SpaceX's recent earnings call, Musk called Grok Bot the fastest-growing product he'd ever seen and said usage per user rises over time instead of falling.
Well, no kidding that growth is off the charts – because he's pricing the product well below the cost to provide it! It'll be interesting to see what happens to growth now that he has restricted usage and will likely raise prices.
This reminds me of how restaurant chain Red Lobster went bankrupt, as this WSJ article recounts:
Red Lobster has run versions of its Endless Shrimp promotion for more than 20 years, and it was a hit with customers. It is also infamous for helping to drive the storied seafood chain into bankruptcy two years ago, when Red Lobster's former owners ran the deal continuously and cratered the chain's profits.
3) Used-car retailer Carvana (CVNA), one of my "Stinky Seven" stocks to avoid, reported third-quarter earnings last Wednesday (press release here and shareholder letter here). The stock jumped 15.7% the next day.
Investors liked the 42% revenue growth to $8 billion, above expectations of $7.7 billion. But the company appears to be cutting prices to drive growth and gain share, as gross profit per unit declined by 5% YOY.
As a result, adjusted earnings before interest, taxes, depreciation, and amortization ("EBITDA") margin fell from 11.3% to 10%. And earnings per share ("EPS") came in at $0.39, below estimates of $0.44.
For insight, I turned to my friend who runs a financial-focused hedge fund. His largest long position is Willis Lease Finance (WLFC) and his largest short is Carvana. He e-mailed me:
I would say that top-line revenues were undeniably good and volumes rebounded a bit versus the second quarter, which is what the market focused on. The question is what Carvana is doing to generate this growth, particularly given that Credit Acceptance Corporation's (CACC) earnings report showed substantial credit deterioration across all recent vintages (2022 to the present), and CACC's stock is trading down nearly 20% on these trends.
Carvana's EBITDA margin was notably softer in the third quarter, and a typically very promotional management will no longer give guidance. They're also changing their reporting, focusing on generally accepted accounting principles ("GAAP") versus adjusted numbers, which I believe is because the "adjustments" they have historically made are becoming smaller and less helpful to operating metrics.
It was a better quarter than the second quarter, to be sure – but the stock's move higher was very surprising, and it is equally surprising that investors don't seem to care about the significant negative signal that CACC's numbers show.
I've warned readers about Carvana more than two dozen times (archive here). And on December 12, 2025, I added it to my Stinky Six – which became the Stinky Seven when I added CoreWeave (CRWV) on August 24 this year.
Since December 12, Carvana is down 21% versus a 10% gain in the S&P 500 Index. It remains one of my least favorite stocks.
Best regards,
Whitney
P.S. I welcome your feedback – send me an e-mail by clicking here.
P.P.S. I posted three photo collages on my Facebook page on Sunday for my 60th birthday – one of baby pictures, one from the last 59 years, and one from the past year:
