Latest / The Bear Market Podcast with Fexingo: Surviving Downturns, Buying the Dip, and Long-Term Resilience / Why Steve Eisman Is Selling AI Stocks Now
Transcript
- Lucas: Steve Eisman — the guy who made a fortune betting against subprime mortgages in the mid-2000s — just sold a major AI stock and went on record saying he's 'starting to have doubts.' That's a pretty stark shift from someone who's been relatively quiet on the AI trade until now. Luna: He sold something, but the headline didn't name the stock. Do we know which one? Lucas: Not explicitly, but given his previous holdings and the context — he's been a holder of NVIDIA and a few other AI plays — it's almost certainly a core AI name. He said he's been trimming position this quarter. And he's not alone: the Moody's warning last week about AI capex threatening credit quality at Amazon, Meta, and Alphabet adds a layer of caution. Luna: Let's start with Eisman's track record. He's not a permabear. He called the financial crisis, but he also missed or stayed out of a lot of big tech rallies. Lucas: Right. That's what makes this interesting. He's not someone who just yells 'bubble' all day. In 2020 he was actually bullish on banks and energy. He's a value-oriented, thesis-driven investor. So when he starts selling a stock he's owned for years, it's worth listening. Luna: And the timing — we've got the VIX at 19.15, up 12% in five days. The Nasdaq dropped nearly 4% this week. The AI trade is getting hit hard. Lucas: Yeah, the S&P 500 is only down 1.5% in five days, but the Nasdaq divergence is real. That's a 2.3 percentage point gap. Eisman said he's worried about the 'magnitude of spending' on AI without clear returns. And Moody's quantified that: they said AI capex could be $50 billion higher than depreciation for the big five tech companies by 2027. Luna: Fifty billion dollars that isn't generating cash flow yet. That's a lot of faith baked in. Lucas: Exactly. And when you add the 30-year Treasury at 5.17%, the cost of capital for that capex goes up. If AI doesn't start showing concrete revenue growth soon, those returns on invested capital could be negative for years. Eisman is basically saying the market is pricing in perfection that may not arrive. Luna: But isn't this just the usual 'this time is different' versus 'it's different this time' debate? Plenty of people sold tech in 2017 and missed the next five years. Lucas: That's the counterargument, and Eisman acknowledges it. But he draws a parallel to the internet bubble — not the technology itself, but the capital allocation. In the late 90s, companies spent billions on fiber optic cable and data centers that took years to monetize. Some winners emerged, but hundreds of billions were destroyed. Luna: So he's not saying AI is worthless, he's saying the capex run-up is priced for a sure thing that isn't a sure thing. Lucas: Right. And now look at the broader market: the VIX is up, the 10-year minus 2-year spread is at 0.36, positive but shallow. That's a steepening yield curve, but it's not screaming 'risk on'. The small caps — the Russell 2000 — are down 1.4% in five days. So it's not a rotation into value either. It's a tech-led selloff. Luna: What about the CME launching single stock futures for SpaceX and others? That seems like an event that says 'the market wants more access to high-growth names.' Lucas: It does, but it's also a sign that traditional IPO markets are still broken for high-profile companies. The hot money wants to trade these names 23 hours a day. It's speculative appetite, not fundamental conviction. That can amplify the selloff when sentiment turns. Luna: So Eisman's sell signal — is it a leading indicator for a broader AI correction, or just one guy taking profits? Lucas: Probably a bit of both. But think about his process: he's a short seller by training. When he sells, he's not just 'taking profits' — he's saying the risk-reward has shifted. And he's doing it while Moody's is downgrading the credit outlook for some of the biggest AI spenders. That's a powerful combination. Luna: We should also look at the VIX itself. At 19, it's elevated but not panic. But the VIX futures curve is in backwardation — near-term higher than longer-term. That means hedgers are paying up for protection right now. Lucas: Yes, VVIX at 101.68, up 5.5% in five days. The market is pricing a higher probability of tail risks. It's not a crash signal yet, but it's a warning. Eisman's move fits into that narrative of 'things are getting frothy and I don't want to hold the bag'. Luna: If our money conversations have helped you think about a decision differently — like when to sell a stock or whether to hold through volatility — that's exactly what listener support helps us keep doing. Lucas: It's a small thing, but if today felt worth a coffee to you, that link is buy me a coffee dot com slash fexingo. No pressure, just a way to keep the show ad-free and independent. Luna: Yeah, we really appreciate it. Now, back to Eisman — let's talk about what he might be buying instead. Lucas: Good question. In the same interview, he mentioned he's adding to positions in financials and energy — traditional value sectors. He specifically said 'if you can get 15% return on equity from a bank at book value, why pay 25 times earnings for an AI stock that might not deliver?'. Luna: So a classic value versus growth rotation. But the data we have — real GDP growth at 2.1%, nominal GDP at $31.86 trillion — that's not a recession. It's a slowing economy. Banks might not be booming either. Lucas: No, but they're cheap. And in a rising yield environment — the 30-year at 5.17% — banks can earn net interest margins. Eisman's bet is that the economy avoids a hard landing and that value stocks catch up. That's a macro call, not just a tech call. Luna: One more thing — the headline also mentioned that Eisman sold 'a key tech stock'. He didn't sell all his AI. He trimmed. That's nuance. Lucas: Important nuance. He's not all out. He's reducing exposure. That mirrors what we've seen from other smart money — like David Tepper reducing his big tech allocation last quarter. The consensus may be shifting from 'buy AI at any price' to 'let's see some evidence first'. Luna: And with the VIX at 19 and VVIX above 100, the options market is already pricing in more downside. So Eisman's move aligns with the broader volatility regime. Lucas: Exactly. When a 'Big Short' investor starts taking chips off the table in the most crowded trade, it's worth paying attention. Not because he's infallible, but because his process forces him to think about where the risk is. And right now, he sees more risk in AI than reward. Luna: So what does this mean for the average listener holding an ai heavy portfolio? Lucas: It might be time to do your own stress test. If NVIDIA or Microsoft or Meta drops 30% — can you hold? If not, maybe trim some on strength. You don't have to go all the way to cash, but Eisman's approach of gradually reducing exposure as the thesis becomes less certain is a discipline worth borrowing. Luna: And if you want to follow the data, the Moody's report is publicly available. Read their assumptions about depreciation and free cash flow. That's the kind of homework that separates investors from speculators. Lucas: Right. We'll keep watching this story. If more big-name investors start selling AI, that's a signal the tide is turning. For now, Eisman's move is a yellow flag, not a red one. But it's enough to make you stop and think.