Steve Eisman: This is How the AI Narrative Collapses
Steve Eisman warns that the AI narrative is vulnerable to a major correction due to three interconnected dominoes: LLM providers lacking competitive moats face pricing pressure from cheaper Chinese alternatives, hyperscalers have massive revenue backlogs dependent on OpenAI and Anthropic's continued spending, and the entire market is overexposed to AI as a single trade. He recommends monitoring the financial health of these LLM providers and favors hyperscalers over pure-play LLM companies.
Summary
Steve Eisman, a renowned investor known for spotting bubbles, expresses serious concerns about the sustainability of the AI investment narrative. He identifies three critical vulnerabilities in the AI ecosystem.
The first domino involves the lack of competitive moats protecting LLM providers like OpenAI and Anthropic. Unlike hyperscalers (Google, Microsoft, Amazon, Oracle) that build the data center infrastructure, LLM providers create the AI models that run on these facilities. Eisman notes that enterprises are increasingly switching between models and adopting cheaper Chinese alternatives like DeepSeek. The cost differential is substantial: DeepSeek V4 Flash costs approximately 3 cents per test compared to 86 cents for Moonshot AI's Kimmy K3, $1.86 for OpenAI's ChatGPT 5.6, and $3.15 for Claude Fable 5. This pricing pressure could trigger a destructive price war.
The second domino concerns the interconnectedness of the ecosystem. A significant portion of hyperscalers' backlogs comes directly from LLM providers: Oracle has roughly $300 billion in backlog from OpenAI alone (out of $640 billion total), and Microsoft announced that 45% of their backlog (approximately $280 billion) derives from OpenAI. Anthropic maintains substantial spending commitments with Amazon and Google. If LLM providers encounter financial difficulties due to pricing wars or demand shortfalls, the hyperscalers would face severe revenue disruptions.
The third domino is systemic concentration risk. Eisman emphasizes that the entire market is essentially "one trade." Even investors believing they are diversified through 60/40 stock-bond portfolios are largely exposed to AI and tech: over 50% of equity allocations are tech and AI-related, and approximately 10-15% of new US corporate bond issuances (roughly $400 billion) are tied to tech and AI companies. This concentration means a significant correction in AI would reverberate throughout the broader market regardless of investment strategy.
Eisman identifies the financial health of OpenAI and Anthropic as the critical indicator to monitor for signs of sustained sector weakness. He anticipates their upcoming IPOs will provide transparency through SEC filings.
Regarding investment positioning, Eisman personally reduced his AI exposure by selling Google. While acknowledging the uncertainty surrounding long-term AI winners, he expresses greater confidence in hyperscalers due to their established business moats, diversified revenue streams, and financial capacity to sustain massive capital expenditures through uncertain times. He is significantly more skeptical of pure-play LLM providers lacking diversification and revenue stability.
Key Insights
- LLM providers like OpenAI and Anthropic lack sustainable competitive moats and face intense competition from drastically cheaper Chinese providers like DeepSeek, which could trigger a destructive price war that threatens the viability of these businesses
- Hyperscalers' revenue backlogs are heavily concentrated in LLM providers, with OpenAI alone representing half of Oracle's $640 billion backlog ($300 billion) and 45% of Microsoft's backlog ($280 billion), creating a systemic vulnerability where LLM provider problems cascade to hyperscalers
- The AI investment narrative represents a single unified trade across the entire market, with over 50% of equity allocations in diversified portfolios being tech and AI-related, and even corporate bond issuances heavily concentrated in AI-related companies, making traditional diversification ineffective
- Hyperscalers possess genuine competitive moats through the massive capital requirements and expenditure thresholds needed to build data center infrastructure, which acts as a sustainable barrier to entry despite their capital-intensive nature
- Google and Microsoft have structural advantages over pure-play LLM providers because they maintain diversified, established revenue streams and hyperscaler businesses that balance vulnerability and provide financial strength to sustain high capital expenditure through uncertain periods
Topics
Transcript
[0:00] What happens if AI doesn't succeed? >> I think we have a big correction. It's all one trade. >> Over the past few years, the investing world has been pretty crazy about the AI companies. So much so the big LLM providers like OpenAI and Anthropic are now looking to cash in with big IPOs. But is the AI narrative actually starting to crack? Well, Steve Eisman, an incredibly smart investor and one who is particularly talented at spotting bubbles, certainly seems to think so. So much so he's getting out of his AI plays. >> I sold my Google. I wanted to reduce my [0:32] exposure to AI. What scares me is that it's all one trade. So…
Full transcript available for MurmurCast members
Sign Up to AccessMore from New Money
Mohnish Pabrai's Honest Thoughts on the S&P500...
Mohnish Pabrai discusses his investment philosophy in the current market environment, focusing on the overvaluation of the S&P 500 and the importance of investing in companies with strong management. He emphasizes the need for investors to identify undervalued, less popular opportunities and to approach uncertainty with caution.
Warren Buffett: "I Initiated Berkshire's Investment in Google"
Warren Buffett revealed he personally initiated Berkshire Hathaway's $31 billion investment in Alphabet/Google, basing the decision on Google's high return on capital, long-term compounding track record, strong competitive moat, and fortress balance sheet. However, Buffett downplayed the investment relative to his other holdings, suggesting it may be primarily motivated by returns exceeding US Treasury yields rather than being a transformative bet.
How Overvalued is the Stock Market, Really?
The video analyzes four key stock market valuation metrics to determine if the market is overvalued. While valuations are historically expensive by most measures, the speaker concludes that a major crash is unlikely in the near term, as today's market differs significantly from the dot-com bubble despite some similarities.
Michael Burry Just Made a Big New Bet...
Michael Burry has invested in Lululemon despite the stock falling 75% from highs due to slowing growth, tariff pressures, and management missteps. Burry argues these are temporary execution problems rather than permanent business deterioration, pointing to the company's history of recovering from past crises and its continued strong cash generation as evidence of a potential turnaround opportunity with expected 18% annual returns.
Howard Marks: AI's IPO Euphoria is a Warning Sign
Howard Marks warns that current AI company IPOs exhibit bubble characteristics similar to previous tech booms, with valuations disconnected from predictable earnings. He argues investors cannot confidently forecast AI company profitability 10 years out, and recommends a spectrum approach: hyperscalers with established moats as lower-risk, pure-play AI companies as medium-risk, and startups as lottery-like high-risk plays.