Michael Burry says he's shorting Caterpillar for the first time after it nearly doubled in the AI-driven rally of 2026

Michael Burry initiated a bearish position in Caterpillar, shorting shares at $1,060.98 after the stock surged 86% in the first half of 2026 and hit record highs. He flagged valuation risk, citing Caterpillar’s price-to-sales ratio at the highest level in at least three decades, and warned AI-infrastructure beneficiaries are becoming overextended. Burry also added bearish bets in Nvidia, Applied Materials, Tesla, and SOXX, noting the Philadelphia Semiconductor Index is ~65% above its 200-day moving average—only seen during the dot-com bubble in 2000—suggesting the rally’s “beginning of the end.”

Analysis

CAT is the cleanest expression of a crowded narrative, not the cleanest expression of AI spend. If the market is paying for hyperscaler capex, the first claims usually accrue to chips, power management, cooling, and grid gear; earthmoving equipment only gets a second-order, slower beta and tends to mean-revert when the growth story shifts from order announcements to delivery and margin discipline.

The near-term risk is positioning, not fundamentals: a celebrity short disclosure can trigger fast de-rating if holders are already leaning on momentum and the stock is priced for perfection. The 1-3 month catalyst path is earnings/guidance, especially whether backlog converts as advertised and whether management has to explain why AI-related demand is not translating into a step-up in incremental margin. If that conversion rate disappoints, the multiple compresses faster than the revenue base.

Contrarianly, Burry may be early on the timing even if he is directionally right. CAT can remain expensive while industrial capex, replacement demand, and power-generation spending stay firm, and a squeeze is likely if the company simply reiterates its order book. The broader short basket is more vulnerable if semis keep beating, so the best falsifier is a continuation of strong guide-ups and SOXX/NVDA holding above trend into the next earnings cycle.

Key Decisions for Investors

• Tactically fade CAT strength with a 1-3 month put spread or short on failed post-earnings/rally extension; target 15-20% downside if the valuation premium starts compressing, and cut if management raises backlog/margin guidance again.
• Prefer a relative-value short: short CAT / long ETN or VRT for 3-6 months to isolate the AI-infrastructure mix away from the ‘AI proxy’ multiple; thesis breaks if CAT shows materially better order conversion than power/cooling peers.
• Keep any bearish exposure in NVDA/AMAT/SOXX modest and event-driven, not outright crowded-beta size; use them only if semiconductor guidance rolls over, with the 200-day moving average as a trend-reversal alert.
• Watch CAT earnings for backlog quality, not headline revenue: if AI-linked orders are real, backlog should convert into margin expansion; if it does not, increase the short on any rally.
• If CAT sells off hard while ETN/PWR/VRT hold up, rotate longs toward electrical infrastructure rather than equipment names; that is the cleaner way to own AI capex over the next 6-18 months.