Intro
Steve Eisman, who ran FrontPoint Partners through the subprime crisis and now hosts The Real Eisman Playbook, tells David Lin that US growth and the AI trade have become the same position, and walks through the revenue lines he will read in the second half to see whether it is holding. Lin brings Congressional Budget Office deficit charts, a headline CPI chart and the case for Treasury intervention to it.
Guest: Steve Eisman, host of The Real Eisman Playbook, formerly of FrontPoint Partners, portrayed as Mark Baum in The Big Short
Host: David Lin
Published: 28 August 2026 on The David Lin Report
Episode page | 32 min
Key Takeaways
The AI food chain and the US economy are now the same bet on two companies that lose money
"I think it's a little nerve-wracking that the entire US economy is dependent upon two companies that lose billions"
Half of this year's roughly 2% GDP growth is AI capex, on his read
OpenAI is the weaker of the two, and Eisman thinks it is in trouble
Revenue up only 18% on the March quarter while costs rose $3 billion in three months
Token maxing ended around late June and July, which makes the second half the real test
"I think the whole tech space sells off" if B2B sales slow
Revenue growth, not profitability, is the metric that breaks the trade
"I think eventually profitability matters, but near-term these stocks don't care"
A Treasury attempt to cap long-term rates is not worth pricing
"that's not going to happen. It's a joke." — $4 billion is not even a basis point
Nvidia's announced $500 billion third-party capital facility has not actually been signed
Embedded gains, not conviction, are why almost nobody de-risks
A 35%-plus capital gains bill on a big Nvidia position is "painful"
Forty years of deficit doom has been wrong because there is no alternative to Treasuries
The repo market is about $3 trillion and it is all T-bills
He is still long Nvidia and several other tech stocks, less long than he was, and not short
The banking system has never been this safe in anyone's lifetime
Housing is where inflation actually bit, through taxes, insurance and utilities
He is not redoing the 2006 subprime work, because the subprime mortgages are not there
From The Big Short to a Breakfast With Steve Carell
Lin introduced Eisman as the investor Michael Lewis wrote into The Big Short, renamed Mark Baum for the book and the film and played on screen by Steve Carell
The show's own framing of the record: Lin said FrontPoint Partners' position doubled from $700 million to $1.5 billion on being right about subprime in 2008
Eisman met Carell exactly once, for about an hour, at a diner near his home called Three Guys: "he wore a Boston Red Sox hat so that nobody would know who he was"
He took Carell home to meet his family, and that was the last substantive conversation the two had
On one breakfast turning into a screen performance: "somehow his portrayal of me was quite accurate"
The Crisis Eisman Says Is Not Hidden
Asked what he would tell Congress about a looming crisis, Eisman said there is nothing hidden about it: US GDP growth this year is going to be up around 2%, and half of that is AI capex
"it's not like I discovered the Rosetta Stone" — he says all of it is public information
Nvidia had just reported revenue up more than 100%, but the line he pulled out was the receivables: 70% of accounts receivable came from five companies, and he said nobody outside knows who they are, though you could guess a bunch of them
The chain as he draws it runs Nvidia to hyperscalers to two customers: 70% of hyperscaler AI revenue is from Anthropic and OpenAI, equating to about 25 to 35% of their cloud revenue
On Oracle alone, he said 50% of its $600 billion backlog is just from OpenAI
"the entire AI ecosystem food chain is dependent upon the future health and success of Anthropic and OpenAI" — and, he added, so does the US economy
"I think it's a little nerve-wracking that the entire US economy is dependent upon two companies that lose billions"
Between the two he called OpenAI "the weak sister", and said prediction markets already have Anthropic reaching an IPO first, at 93%
Token Maxing Ended, and the Second Half Is the Test
Eisman expects Anthropic's revenue story to be good through June, and dates the change to around late June and July, when what he calls token maxing ended
Before that, he said, people had no sensitivity to spending and just spent money like crazy on tokens; now they are a lot more self-conscious
He thinks the third and fourth quarters will be much more interesting for Anthropic and OpenAI than the first half of the year
He is hoping for some clarity when Anthropic puts out its S-1, though "Whether they'll give it or not, I don't know"
Asked what slowing business sales would mean for semiconductors and the rest of technology: "I think the whole tech space sells off"
Lin asked whether July's selloff was already that. Eisman: "That was like the preview"
Lin pressed for the metrics that would settle it, noting that "the word bubble is hotly debated"
The Metrics: Second-Half Revenue Growth, Not Profits
The variable he named is revenue growth for the second half of this year
On the June quarter as he recalled it, Eisman put Anthropic's revenue at about $11.5 billion, "up over 100% versus the March quarter"
OpenAI, which he called "the problem child", had around $6.5 billion in revenue and was up only 18% on the March quarter
Worse, on his numbers, its costs went to $12.5 billion, up $3 billion in three months — revenue up a billion sequentially against costs up three: "That's not the right direction."
Lin put Tesla to him, unprofitable for most of its history while the stock went up because investors believed in a bigger picture. Eisman: "I think eventually profitability matters, but near-term these stocks don't care"
He finished the thought across Lin's interruption — revenue growth slowing "is the more important metric"
Why He Calls a Treasury Rate Cap a Joke
Lin began a question about the Treasury successfully capping long-term rates. Eisman cut in: "that's not going to happen. It's a joke."
On the scale of it, he said the Treasury has $4 billion to buy Treasuries with — "It's not even a basis point. It's silly."
Lin said people agree with him but argue the signaling is what matters. Eisman's answer was that a $4 billion signal is not one anybody should care about
On the maximalist version, he said Scott Bessent could point at a Treasury general account holding a trillion, but "He's not going to spend a trillion dollars"
He called the idea that Treasury should spend a trillion dollars to try and cap long-term rates absurd, and said quantitative easing already ran the experiment: "It had absolutely zero impact on the US economy. The only impact it had was causing stock prices to go up."
Asked whether 5% is the cap on the 10-year Treasury yield, he declined the question entirely: "I never make predictions about interest rates", with Lin agreeing that "no one can predict interest rates"
Nvidia's $500 Billion That Doesn't Exist Yet
Asked whether low rates and available capital keep the train running, Eisman said no: "First of all, OpenAI I think is in trouble."
Its revenue growth has obviously slowed dramatically, it had better start accelerating soon, and with token maxing over he thinks it will be harder for these companies to raise capital — "I could be wrong, but that's what I think"
Lin raised Nvidia's announced $500 billion third-party capital facility with KKR, Apollo and other private-capital firms. Eisman said that is not exactly what happened
On his account, citing Bloomberg, the transaction was taking a long time, so "Jensen went out and basically just announced it anyway"
What exists is closer to a memorandum of quiet understanding: "they may get it done but right now there is no 500 billion dollars"
Lin offered a dinner-party version of the thesis — "it sounds to me like the entire economy is propped up by a couple companies that may or may not even remain profitable going forward" — and asked if it was true
Eisman's answer was one line: "I think that's 100% accurate."
Why Nobody Sells: Embedded Gains and a 35% Tax Bill
Asked how an ordinary investor prepares for this, he said it is very very difficult, and that the obstacle is unrealized gains held by institutions and individuals alike
Working through a hypothetical holder who bought Nvidia well and made many times their money: "If you sell your Nvidia, you're going to pay 35% plus in capital gains taxes. That is painful."
He has no great answer for people carrying embedded gains, and thinks most of them simply will not sell — which is what makes the question so difficult
The one concrete lower-risk instrument he named was the KBWP, the property casualty insurance index ETF, offered as just one example
The Deficit Argument That Has Been Wrong for 40 Years
Lin shared Congressional Budget Office projections showing federal debt as a share of GDP continuing to rise, possibly past the World War II high, with the deficit widening on the same path
Eisman said all other things being equal a smaller deficit would be better, but he is not of the view that the US economy crashes and burns anytime soon because of it
He put today's doom case in a line going back to Pete Peterson moaning and groaning about the deficit in the 90s, and said the reserve-currency-collapse argument has been made for 40 years
"when you make an argument for 40 years and you've been wrong for 40 years, probably you should ask yourself like, okay, why hasn't what I predicted happened?"
His answer is liquidity: the US Treasury market is the most liquid bond market in the world, and so "the financial system of planet earth uses Treasuries"
His illustration is the repo market, where banks lend to one another overnight — he put it at about $3 trillion, all T-bills and nothing else
Asked why an alternative is not seriously considered, he said there is nothing to consider: not Chinese bonds, not Bitcoin, not European bonds, because "for there to be an alternative, you need to be as big and as liquid"
On Ray Dalio, whom Lin raised as the source of the argument: "Ray Dalio has been talking about this nonsense for the last 20 years"
The caveat he does accept is rates, not calamity: the amount of debt being raised by AI is crowding out everything else, so rates could go higher, which would hurt housing and the economy
Gold, Bitcoin, Burry, and a Man Who Won't Time Markets
Lin described a market that turned positive from around 19 August: Bitcoin reawakening along with most of crypto, and gold moving from $4,000 to $4,700 on what people believe is Treasury intervention
Eisman would not call it: "I have literally no idea. You know, I just don't like to predict markets."
The risk he keeps returning to is Anthropic and OpenAI, which he calls huge — but he said that if he is right it could materialize a year from now, and "market timing is everything"
Lin noted Michael Burry is betting against Nvidia and shorting the chip sector, and asked whether Eisman would go short. He was explicit about his own book: "I am still long Nvidia and several other tech stocks. I'm less long cuz I've gotten nervous, but I am not short."
He might hedge those longs — "I might. I might. It's possible." — but on what the hedges would be: "I'd rather keep that to myself for now"
The K-Shaped Economy and the Single Trade Underneath It
The Iran war does not change his economic thesis: he thinks it has long-term implications for the United States, but that the economy as of now is fine
"we do have a K-shaped economy. The bottom of the K is struggling. There's no question about that" — and yet, as long as AI capex keeps powering higher, he said the economy and the market probably go higher
"It's all it's all one trade."
Lin asked whether the stock market has to be propped up or everything falls apart, and suggested that is one argument for why the Treasury is intervening ahead of the midterm elections. Eisman: "I think that's fair."
He then gave the conditional the episode is named for, with the disclaimer attached in front of it — "and I'm not making this prediction, but I'm just saying if tomorrow" — before finishing: "If tomorrow OpenAI failed, the US economy, I think, would go into an immediate recession and the market would have a massive correction."
He does not think Treasury action changes that outcome either way: "OpenAI will succeed or fail because it'll succeed or fail."
Chinese Open-Weight Models and the Price War He Expects
Lin recalled Eisman telling CNBC that Chinese open weight models are much cheaper, and put it to him that they are starting to take share
Asked why US buyers are not using them at scale yet, Eisman said he is hearing that people are starting, and that it takes time
If that shift does happen, his expectation is blunt: "there would be a price war"
What His Own Guests Have Been Telling Him
Asked what the guests on The Real Eisman Playbook have been worried about, Eisman started with Wolfgang Münchau, author of Kaput: The End of the German Miracle, on why Europe does not grow
"regulation in Europe basically kills the tech sector in Europe" — Germany, France and the rest, he said, literally don't grow
After Meta settled a California lawsuit over its algorithms, he had on Fordham law professor Benjamin Zipursky to explain the legal theories behind those suits, which he called fascinating
Looking for something to mark the 250th anniversary of the United States, he found Capitalism in America: An Economic History of the United States and interviewed the co-author Adrian Wooldridge, since he could not interview Alan Greenspan
He describes the show as very eclectic — sell-side analysts, authors and economists — and says a listener gets a real education from it
The 1905 Trade That Took Down Situational Awareness
Lin raised Situational Awareness, and the view that it was directionally right and badly risk-managed. Eisman agreed in one word: "Totally."
His analogy, from one of his weekly wraps: "Imagine it's like 1905." An investor convinced that automobiles will conquer the world and bankrupt the carriage companies buys every public auto and auto parts company and shorts every horse carriage company
The trap is that the longs and shorts are not independent of one another — they move exactly in unison, so the whole book is a single trade, even though he says the underlying thesis would prove 100% correct over the long term
The fund's version was long hardware and short software; a terrible car accident that gets a tremendous amount of press reverses the trade for a period while people wonder whether cars are safe
"if you're levered 4 to 1, you get put out of business really quickly. That's what happened to him"
Why the Banks Don't Worry Him
Asked whether the financial world has learned the lessons of 2008, Eisman went to leverage: "the banks post Dodd-Frank were forced to delever enormously", to something like half of what it was
They may have exposure to private credit, he said, but given how much capital they hold he does not think it is a systemic issue
Lin asked whether that counted as a lesson learned or something rammed down their throats. Eisman: probably a combination of the two
"the banking system in the United States, I don't think has ever been this safe in anyone's lifetime"
Where Inflation Actually Bites: Housing's Hidden Costs
Lin showed a headline CPI chart and argued that worrying about consumer prices is a recent phenomenon by the standards of the last 20 years: CPI averaged between 1 and 2% after 2008, with a peak in 2011 that came back down and a decade in which people did not have to think about it
Eisman's answer was housing: "I think the area that's been really really hurt is housing"
The hidden costs he lists are taxes, property and casualty insurance, and utilities
"there's a whole swath of Americans who have been completely priced out of the housing market" — leaving aside the price of the house, he said the annual costs alone are beyond their means
He thinks that shows up in housing-related stocks and not much beyond them
Lin's last question was what would make him redo the 2006 and 2007 work, when he dug into California and Florida subprime mortgage data to build a thesis that housing was in trouble. Eisman said there really are no subprime mortgages now
"If there was a correction in the housing market I think it would just be a normal correction." — not calamitous, and not where he puts his attention these days
Eisman's bottom line is that the AI trade and the US economy have become one position, and that what settles it is not the deficit, the Federal Reserve or the Treasury but whether Anthropic's and OpenAI's revenue growth holds up in the second half.
Products, Companies & Tools Mentioned
OpenAI and Anthropic (The two loss-making companies Eisman says the entire AI food chain and the US economy now depend on; he rates OpenAI the weaker of the pair and says it is in trouble)
Nvidia (Just reported revenue up more than 100%, with 70% of accounts receivable from five unnamed companies; also the stock Eisman is still long, Michael Burry is short, and his hypothetical investor cannot sell for tax reasons)
Oracle (Eisman says 50% of its $600 billion backlog is from OpenAI alone)
KKR and Apollo (Named by Lin among the private-capital firms in Nvidia's announced $500 billion third-party facility, which Eisman says has not actually been done)
Tesla (Lin's counterexample — unprofitable for most of its history while the stock went up on a bigger picture)
Meta (Settled a California lawsuit over its algorithms, which prompted an episode of Eisman's show on the legal theories behind it)
Situational Awareness (The fund that blew up running long hardware against short software at 4-to-1 leverage)
KBWP (The property casualty insurance index ETF, Eisman's one named lower-risk example for nervous holders)
FrontPoint Partners (The fund Eisman ran in 2008; Lin said its position doubled from $700 million to $1.5 billion)
Books & Resources Mentioned
The Big Short – Michael Lewis (The book that made Eisman famous, and the film in which Steve Carell played his renamed character)
Kaput: The End of the German Miracle – Wolfgang Münchau (Münchau came on Eisman's show to explain why Europe does not grow)
Capitalism in America: An Economic History of the United States – Alan Greenspan and Adrian Wooldridge (Eisman interviewed Wooldridge for the 250th anniversary of the United States)
The Real Eisman Playbook (Eisman's own show, where the weekly wrap carried his 1905 analogy and where the Münchau, Zipursky and Wooldridge interviews ran)
Get the latest market chatter and takes as they happen:
X | Threads | Instagram | YouTube | TikTok | Facebook