Chinese memory ban would cut off RAMpocalypse relief

US lawmakers urged the Trump administration to tighten export controls on Chinese memory makers YMTC (NAND) and CXMT (DRAM), warning that US purchases would subsidize the PLA and undermine Western suppliers. They asked for an executive order/directive to block US persons and US-incorporated firms from procuring chips from entities on the BIS Entity List or DoD Section 1260H list (notably CXMT and YMTC). The article notes memory prices are expected to stay high through at least 2028, while OEMs (e.g., Apple, Dell, HP) are reportedly beginning to qualify Chinese chips amid shortages.

Analysis

The cleanest read is not “more sanctions” but a longer-for-longer pricing regime. If US buyers are blocked from a lower-cost, qualified alternate source, the marginal pricing power stays with Micron/SK Hynix/Samsung, which extends gross-margin support into the next several quarters rather than solving the shortage. That is bullish for memory suppliers, but the bigger market reaction should be on OEMs and device assemblers that cannot pass through memory inflation fast enough: AAPL, DELL, and HPQ face a margin squeeze first, then a potential demand hit if configs are held back to protect ASPs.

Second-order, this is a substitution problem, not just a supply problem. If Chinese parts are removed from procurement plans, buyers will absorb more cost from incumbents until new wafer capacity arrives; that likely keeps per-unit memory spend elevated through 2026 even if shipment volumes normalize. The risk for the long memory trade is that the market already prices a very good cycle: once capacity additions from Micron/SK Hynix hit in volume, the incremental upside from policy-tightened Chinese competition fades and the multiple can compress quickly.

Contrarian view: the consensus is likely overstating how much this changes near-term supply and understating how much it changes negotiation leverage. Most OEMs can qualify alternatives, but they cannot re-engineer BOMs instantly, so the immediate loser is gross margin, not units. The key falsifier is an actual executive order or agency directive; without that, this stays a headline risk, not a fundamental reset.

Key Decisions for Investors

• Short AAPL/DELL/HPQ basket for 4-8 weeks: highest risk/reward is on OEM gross-margin compression if memory ASPs stay elevated; cover on any indication procurement can source alternative supply without price concessions.
• Long SKHYV / SYBJF into any policy headline fade for 1-3 months: incumbent memory makers should keep pricing power if Chinese supply is restricted, but trim if spot DRAM/NAND prices roll over or if management commentary signals demand destruction.
• Pair trade: long memory producers vs. short PC hardware ETF / OEM basket (e.g., MU or SKHYV vs. AAPL/DELL/HPQ) for the next earnings cycle; thesis breaks if OEMs guide to stable gross margins despite memory inflation.
• Watch for a real policy catalyst before sizing up the short: no action on an executive order or BIS directive means the trade remains tactical, not structural.