Back to News
Market Impact: 0.28

Michael Burry launches newsletter to lay out his AI bubble views after deregistering hedge fund

NVDA
Artificial IntelligenceTechnology & InnovationInvestor Sentiment & PositioningMonetary PolicyShort Interest & ActivismCompany Fundamentals
Michael Burry launches newsletter to lay out his AI bubble views after deregistering hedge fund

Michael Burry has deregistered his hedge fund and launched a paid Substack, 'Cassandra Unchained' ($379/year), to set out a bearish thesis that the current AI-driven market resembles the dot‑com and pre‑2008 housing bubbles. Leveraging 1.6 million X followers, Burry cited Fed commentary (Powell) and Greenspan-era echoes to argue investors are overvaluing AI firms, and has been openly bearish on high‑profile AI names such as Nvidia and Palantir — a view that may weigh on sentiment around richly valued technology companies.

Analysis

Market structure: Sentiment-driven de-risking will favor lower-PE, cash-generative incumbents and bond-like tech cash cows while pressuring high-PE, revenue‑multiple AI plays (NVDA-style) through multiple compression of 10–30% if flows reallocate over 4–12 weeks. Pricing power for core AI suppliers is bifurcating: firms with long-term manufacturing scale (TSM, ASML) retain scarcity rents, while pure‑software/expectations names face volatility and funding withdrawal that can amplify supply/demand mismatches in secondary share markets. Cross-asset: expect S&P options skew to steepen, short-dated IV on NVDA to rise 30–60%, and modest widening in IG spreads as equity risk premium ticks up; USD strength is likely in risk-off episodes. Risk assessment: Tail risks include regulatory export controls or antitrust action that could truncate TAM (mid-tail, 5–15% revenue downside for exposed suppliers) and a crowded short that produces a gamma squeeze (low-probability, high-impact). Immediate (days) effects: IV spikes and retail-driven flow; short-term (weeks–months): re-rating events around earnings and Fed commentary; long-term (quarters–years): fundamentals (capex cycles, wafer capacity) reassert value. Hidden dependencies: enterprise cloud budgets and foundry lead times create 6–18 month lags between sentiment and real revenue, so price moves can overshoot fundamentals. Key catalysts: NVDA quarterly guide, Fed minutes, semiconductor capacity announcements. Trade implications: Tactical plays: buy protective 3‑month put spreads on NVDA 10–20% OTM sized to 1–3% portfolio risk; pair long MSFT (2–4%) vs short NVDA (1–2%) to capture cloud revenue stickiness vs hardware-expectation risk. Use calendar spreads to sell short-dated calls into high IV and buy longer-dated puts to hedge conviction; consider 30–90 day straddles around NVDA earnings only if expecting >25% move. Rotate 3–6% from AI equity exposure into Financials/Industrials ETFs over 6–12 weeks to harvest valuation dispersion. Contrarian angles: Consensus underweights the stickiness of data-center capex and the oligopolistic supply of advanced GPUs, which could rebound pricing power beyond current expectations; if NVDA guidance remains intact, a 6–12 month recovery of 20–40% is plausible. Reaction may be overdone in names with durable enterprise contracts (PLTR faces sentiment risk but only limited immediate revenue hit); conversely, crowded shorting risks include squeezes given concentrated float—size shorts accordingly and prefer option-defined risk. Historical parallel: 2000 tech drawdowns created buying opportunities for infrastructure leaders that later consolidated market share; repeat is possible here if one-layer sellers dominate.