Michael Burry just started a massive group chat on Substack, and it's as chaotic as you'd expect

Michael Burry launched a paid Substack, 'Cassandra Unchained', and opened a paid-subscriber group chat that quickly generated hundreds of replies and helped the newsletter reach more than 97,000 subscribers; he has 1.6 million followers on X. Burry disclosed holding bearish put options on Nvidia and Palantir and has publicly warned of an AI bubble, a combination of public platform and options positioning that could subtly influence sentiment and trading interest in key AI/technology names.

Analysis

Market structure: Burry’s public put disclosure and viral Substack chat amplify retail/short‑interest signaling rather than fundamentals — winners are platforms (Substack/X), options dealers capturing widened IV, and nimble volatility traders; losers are concentrated NVDA and PLTR equity holders and momentum funds that rely on easy risk-on flows. Short-dated put demand will lift implied vol and skew, raising hedging costs for large long-only holders and increasing bid for protection by ~5–15 vol points near spikes.

Risk assessment: Immediate (days) risk is an IV and headline spike with potential for gamma‑squeeze if retail flips to buy calls; short‑term (weeks/months) risk is crowded positioning and forced deleveraging in semis/AI ETFs; long-term (quarters/years) fundamentals (AI revenue growth, product cycles) likely reassert. Tail risks: regulatory action on AI chip monopolies or a sudden guidance pullback from NVDA/PLTR could cascade into correlated equity drawdowns and liquidity stress in single‑name options.

Trade implications: Favor defined‑risk, small‑size hedges rather than outright large shorts — buy put spreads on NVDA/PLTR (3–6 month expiries) and avoid naked shorting into elevated retail attention. Consider pair trades to reduce idiosyncratic risk (short NVDA vs long broader semiconductor exposure) and rotate 2–4% portfolio weight from mega‑cap AI into cyclicals/value to lower crowding.

Contrarian angles: The market may over-index on Burry’s brand rather than position size — a small, well‑publicized put can spike IV but may not reflect earnings power; historical parallels (GME/AMC) show social amplification can produce short squeezes, not linear declines. Unintended consequence: too many players buying protection makes hedges expensive — optimal entry is on IV mean‑reversion or after an earnings/option‑expiry washout.

Key Decisions for Investors

• Trim NVDA exposure by 2–4% of portfolio if overweight; redeploy 1–2% into AMD (ticker: AMD) and 1–2% into SOXX to capture semiconductor breadth and reduce single‑name concentration—execute within the next 10 trading days or ahead of NVDA earnings/catalyst.
• Establish defined‑risk protection: allocate 1–2% portfolio notional to 3‑ to 6‑month NVDA put spreads (buy ~10% OTM put, sell ~20% OTM put) and 0.5–1% notional to equivalent PLTR put spreads to cap premium; close or roll if 30‑day IV falls >10 vol points or stock rallies >25% within 30 days.
• Initiate a pair trade: short 0.5–1% notional NVDA and long 1–1.5% notional SOXX/AMD to express relative weakness in the leader while keeping semiconductor upside; re‑balance after 4–8 weeks or if NVDA reports>15% revenue surprise.
• Monitor concrete signals before scaling: if NVDA weekly put open interest rises >30% WoW or 30‑day IV exceeds its 60th percentile, add to defined‑risk hedges up to the sizes above; conversely, if IV mean‑reverts by >8–10 vol points, trim hedges and redeploy into value/cyclicals.