Michael Burry is shorting Nvidia and Micron, but won't touch this AI stock

Michael Burry is getting more aggressive on the AI trade, keeping bearish exposure via put options on Nvidia and Palantir. He also increased his Micron short and rotated out of underperforming semiconductor ETF puts into a larger bearish position in the Nasdaq-heavy QQQ, signaling continued downside risk in AI-linked semiconductors and broader tech.

Analysis

This looks less like a call on individual product roadmaps and more like a bet that the market is overpaying for AI duration. If that view is right, the first-order losers are the highest-multiple, most narrative-sensitive names in the AI complex — but the bigger second-order damage is to index-level tech beta: once a few crowded holders de-risk, passive flows can force selling across semis, software, and internet, even where fundamentals are fine. That makes QQQ the cleaner expression than a single-name short if the goal is to capture a de-rating event.

The contrarian read is that the AI capex cycle is still early enough that shorts can be structurally wrong on timing. NVDA remains a supply-constrained cash-flow machine as long as hyperscaler budgets hold, and PLTR is more vulnerable to multiple compression than operational collapse; if either prints upside on growth or raises forward guidance, the bear case can unwind quickly. In that sense, the thesis needs a catalyst beyond "AI enthusiasm is crowded" — it needs evidence that enterprise spending is pausing or that hyperscaler capex growth is decelerating.

Over the next 1-3 months, the key risk is a benign earnings season: even modestly positive AI spend commentary could trigger a short squeeze because positioning is already tilted cautious. The intermediate catalyst to watch is memory and semiconductor lead indicators — if inventory, pricing, or backlog data soften, the bear case broadens from sentiment to fundamentals. Falsification is simple: sustained upward revisions in NVDA data-center demand, PLTR commercial acceleration, or a stabilizing/expanding QQQ breadth profile would argue the market is still rewarding earnings power, not just story.

Key Decisions for Investors

• Use QQQ as the primary expression of the bearish AI/multiple view: buy 1-2 month put spreads on strength into earnings, targeting a move lower only if megacap tech breadth deteriorates; risk/reward is better than shorting a single AI winner.
• Stay cautious on outright NVDA shorts until the next guide/earnings catalyst; if initiating, prefer a small-sized bear put spread rather than stock short because any hyperscaler capex reaffirmation can squeeze the name sharply.
• Treat PLTR as a valuation-sensitive momentum short, not a fundamental collapse short: consider a pair trade short PLTR / long a lower-multiple defense software name if growth is resilient but the market starts discounting duration.
• Watch SMH and SOXX relative to QQQ over the next 2-6 weeks; if semis start underperforming while QQQ holds up, that suggests the de-rating is staying contained, which would argue against scaling the bearish trade.