Michael Burry adds to shorts as Nasdaq-100 hits all time high. Why he's betting against chips

Michael Burry warned that the AI-driven memory-chip shortage could reverse within two years as production, particularly from China, catches up, triggering another cyclical memory downturn. He increased put positions in Micron, Nebius, the SOXX semiconductor ETF and Palantir, while calling the Nasdaq-100 historically overvalued and top-heavy after it reached a record 30,732. Concerns are reinforced by rising Chinese DRAM and NAND capacity, which could create surplus supply, pricing pressure and trade friction; Nasdaq-100 breadth has also weakened, with the share of constituents above their 200-day average falling to 56% from 77% since mid-August.

Analysis

The relevant risk is not simply lower memory pricing, but a regime change in MU's earnings-multiple support: AI-driven HBM supply remains differentiated, while legacy DRAM/NAND carries the volume and inventory risk if incremental Chinese supply is redirected into commodity segments. That would first compress spot and contract pricing, then reduce the cash-flow durability embedded in forward estimates; the key 1-3 month indicators are quarterly contract-price resets, customer inventory days, and MU's FY capex-to-revenue guidance. Export controls may keep Chinese suppliers structurally constrained in leading-edge HBM, making a broad 'memory glut' thesis premature unless commodity weakness leaks into high-bandwidth memory pricing or displaces non-AI DRAM demand materially.

The more actionable near-term implication is rising AI dispersion rather than an index-level collapse. MU and higher-duration software beneficiaries such as PLTR are vulnerable because both require sustained elevated growth expectations, whereas SOXX contains foundry, analog, equipment and diversified logic exposure that may not share memory's cycle. A weakening breadth backdrop raises the odds that passive QQQ flows mask deteriorating fundamentals until an earnings miss or guidance reset creates an abrupt de-risking event.

Consensus may be underestimating China's effect on commodity memory but overestimating its ability to substitute for premium AI memory. The bearish thesis is falsified if MU maintains HBM allocation discipline, lifts FY revenue/GM guidance, and reports stable or rising non-HBM contract prices; conversely, a sequential decline in commodity ASPs alongside higher capex commitments would expose negative operating leverage over the following 6-18 months. The cited cash-flow figures should be independently reconciled to reported filings before being used in valuation work, as the magnitude appears inconsistent with conventional forward-FCF framing.

Key Decisions for Investors

• Initiate a 1-3 month relative-value hedge: short MU versus long NVDA or AVGO in equal semiconductor-beta dollars. This isolates commodity-memory/capex risk from AI compute demand; cover if MU raises gross-margin guidance or HBM pricing remains firm at its next earnings update.
• Do not adopt a broad SOXX short solely on this signal. Instead, buy 3-6 month SOXX put spreads only after a break below its 50-day moving average accompanied by worsening semiconductor breadth; target roughly 2:1 payoff and cap premium at 50-75 bps of NAV.
• Reduce tactical PLTR exposure into strength or pair short PLTR against long profitable software/enterprise cash-flow exposure such as MSFT. PLTR's valuation is more exposed to a growth-duration reset than to memory fundamentals; invalidate on material upward revision to FY revenue growth and remaining-performance-obligation acceleration.
• Maintain MELI, SFM, BBW and BIRK as separate idiosyncratic longs rather than treating them as confirmation of a technology short. Re-underwrite each at earnings; their potential diversification benefit disappears if a broad risk-off move raises consumer-credit or discretionary-demand concerns.
• Create an alert for MU: act more aggressively bearish only if commodity DRAM/NAND contract pricing turns negative sequentially for two monthly checks, customer inventory builds, or management raises capex without corresponding revenue guidance. Without these data, the supply thesis is a watch item rather than a high-conviction directional short.