Semis Surge, Michael Burry's Shorts Feel the Pressure — But Don't Expect Him to Run for Cover After His Latest “1987-Style” Crash Warning

The article frames Michael Burry’s bearish put bets against AI/memory/semi names (incl. Micron and Nvidia) as potentially timing-risky, arguing overvalued/overbought stocks can stay elevated longer. It highlights uncertainty around whether a “1987”-style top/crash arrives soon, implying downside could be delayed and volatility cut both ways. It calls Burry’s Oracle short the most questionable, citing its ~16.0x forward P/E despite leverage and a credit downgrade, and suggesting a long setup may be more favorable than shorting at that level.

Analysis

The tradeable mechanism here is positioning, not fundamentals: a crowded AI basket can gap lower if dealers, systematic funds, or retail call buyers de-risk together. That makes the first-order losers the highest-beta narrative names (PLTR, TSLA) and the most crowded semis (NVDA, MU, SOXX/SMH), while the second-order casualty is short-vol exposure that’s been monetizing calm. If the tape turns, the unwind will likely be faster in options-rich names than in lower-beta industrials.

The key catalyst window is 1-3 months, not a same-day call on valuation. A bearish thesis only becomes actionable if it is confirmed by weaker capex guides, slower data-center orders, or a miss on monetization metrics; absent that, multiple compression can persist but not resolve. ORCL is the odd one out: after de-rating, the short is lower quality asymmetry than shorting momentum leaders, because some balance-sheet stress is already visible and the market has partially priced it.

Contrarian view: the consensus may be underestimating how much marginal capital is still sitting on the sidelines for AI winners, which can keep expensive stocks expensive. The better bearish expression is not naked shorting into strength, but defined-risk hedges or relative-value shorts against less-valued cash generators. What would falsify the bear case is another round of upward AI revenue/capex revisions or a post-earnings breakout in NVDA/SMH that holds through the next reporting cycle.

Key Decisions for Investors

• Use 1-3 month put spreads on SMH or QQQ as a hedge, initiated on any post-rally strength; defined risk is preferable because timing the unwind is the whole problem.
• If seeking a relative-value short, pair short PLTR or TSLA against long QQQ/XLK rather than outright shorting NVDA; these names have more valuation air and weaker fundamental anchors.
• Do not short ORCL naked here; if you need expression, prefer a modest long ORCL vs short a more crowded AI basket (SMH or PLTR) because ORCL’s downside already reflects more credit and leverage concern.
• Set a hard stop on bearish AI exposure if NVDA/SMH reclaim and hold recent breakout levels into the next earnings cycle; that would indicate the crowd is still being paid to stay long.
• Watch for a capex-revision tell from hyperscalers over the next 30-60 days; absent lower spend guidance, the bear thesis is more likely to remain a volatility trade than a directional collapse.