Michael Burry has placed a roughly $1 billion bet against AI stocks, publicly calling the sector a 2008-style bubble.1 At the same time, Berkshire Hathaway's cash reserves have hit a record billions — the largest defensive position Warren Buffett has ever held.
Three bearish signals have converged simultaneously: Burry's short position, his published Substack commentary, and Berkshire's record cash pile.1 That kind of alignment among contrarian investors is historically rare and has drawn scrutiny from portfolio managers from London to Tokyo.
Burry's 2008 housing short preceded the sector's collapse by roughly 12 months.1 He is now applying that same framework to AI — a sector that absorbed hundreds of billions in global investment between 2023 and 2025, with valuations increasingly detached from near-term revenue across US, European, and Asian markets alike.
Berkshire's cash accumulation adds a systemic dimension. Buffett's record cash positions have historically foreshadowed wide market dislocations, not just sector corrections.1 If that pattern holds, exposure extends beyond AI stocks to global equity markets broadly.
The timing is sensitive for AI capital markets worldwide. Venture investment in AI infrastructure, model development, and enterprise software surged through 2024 and 2025 — not only in Silicon Valley, but in London, Paris, Singapore, Seoul, and Riyadh. Valuation compression in US public markets typically flows into late-stage private valuations within one to two quarters, tightening funding globally.
The 3-to-6-month correction window implied by this signal cluster coincides with a critical period for AI companies eyeing public listings or secondary rounds across major exchanges.1 Sustained weakness in AI stock prices would reduce appetite for those deals and compress exit multiples for investors on every continent.
Burry and Buffett rarely move in the same direction. When they do, history suggests the combined signal carries weight — even if the precise timing of any correction remains uncertain.
