Michael Burry Predicts This Is When the Memory Shortage Will End

Michael Burry expects the memory shortage to unwind within two years as production catches up with demand, triggering another industry downcycle despite AI-driven optimism. He warns Micron could face a steep correction if investors abandon expectations for shortages lasting through the decade; the stock has risen more than 1,300% over five years but is already down 16% from its earlier 2026 high. Rising Chinese capacity, cited by Acer CEO Jason Chen, adds to concerns that supply conditions are easing.

Analysis

The article’s market-cap assertion is materially inaccurate, which weakens its valuation framing; the actionable issue is nevertheless real: MU’s earnings power remains unusually sensitive to HBM pricing, DRAM mix, and the pace at which legacy DRAM capacity is converted back from high-margin AI products. A two-year supply response is plausible, but it would first pressure commodity DDR/NAND pricing rather than necessarily HBM, where qualification cycles and customer concentration create a more durable supply constraint. The key second-order risk is that Samsung and SK Hynix add HBM capacity while protecting utilization by pushing excess output into conventional DRAM, accelerating a broader pricing downturn.

Near term, this is not a standalone short catalyst: Burry commentary and a retail-media article do not change fundamentals, while AI server builds can keep HBM tight through multiple quarters. Over 1-3 months, monitor monthly DRAM contract-price data, HBM lead-time commentary, and MU/Samsung/SK Hynix capex revisions; weakening conventional DRAM prices alongside rising inventory would challenge forward gross-margin assumptions before reported revenue rolls over. Over 6-18 months, the relevant risk is multiple compression if investors shift from a structural-AI scarcity framework back to a cyclical memory framework. Contrarianly, consensus bearishness on a future downcycle may be premature if HBM3E/HBM4 yields and qualification bottlenecks prevent effective supply from scaling as quickly as wafer-capacity announcements imply.

Key Decisions for Investors

• Do not initiate a directional MU short solely on this article; treat it as an alert to reassess exposure after the next MU earnings call and industry contract-price release. A credible bearish trigger is two consecutive months of falling DRAM contract prices combined with no upward revision to HBM bit shipments or gross-margin guidance.
• For existing MU longs, reduce beta or collar 3-6 month exposure if MU trades materially ahead of earnings revisions; use puts or put spreads rather than an outright short, since HBM supply disruptions or stronger hyperscaler capex can produce sharp upside gaps.
• Express a cyclical-memory normalization view through a relative-value basket: short MU versus long NVDA only after evidence that memory pricing is declining while accelerator demand remains intact. The thesis is that MU’s operating leverage reverses faster than NVDA’s platform economics; exit if MU raises HBM pricing, margin, or capex guidance.
• Watch Samsung Electronics and SK Hynix production/capex disclosures for evidence of capacity conversion rather than headline wafer additions. If HBM supply expands without commensurate AI-server demand, MU downside could emerge over 6-18 months; if qualification delays persist, avoid the short thesis.