President Trump Sells Micron Stock and Buys an AI Stock Up 1,340% Since 2023

Portfolio disclosures show President Trump’s investment accounts were net sellers of Micron (net sales $90,000–$116,000) and net buyers of Nvidia (net purchases $246,000–$3.7M) year-to-date through May. Micron reported Q2 FY2026 revenue up 345% to $41.4B and non-GAAP EPS up 1,200% to $25.11, helped by 16 multiyear supply agreements, though Wall Street expects memory pricing to soften as soon as 2028. Nvidia posted Q1 FY2027 revenue up 85% to $81.6B and non-GAAP EPS up 140% to $1.87, with key catalysts including Vera Rubin volume production in 2H and the RTX Spark PC superchip, while Street expects ~56% annual adjusted earnings growth.

Analysis

This remains a full-stack AI share capture trade, not just a GPU story. NVDA’s edge is that every incremental layer it adds — networking, CPUs, systems, and software — raises switching costs and compresses the addressable budget for AMD, INTC, and even some of the value pool that would otherwise accrue to OEMs. The more customers standardize on one architecture, the more the winner monetizes the entire rack, not a single chip.

For MU, the important shift is that longer-dated contracts reduce near-term earnings beta but also cap the upside of a classic spot-memory squeeze. That makes the stock less of a pure cyclical call and more of a duration asset tied to contract enforcement and customer concentration; if memory pricing rolls over earlier than expected, the market will quickly de-rate the “cheap” multiple. The structural risk window is 6-18 months, but the immediate tape reaction is mostly sentiment, not a clean fundamental read-through.

The contrarian point is that consensus may be underpricing NVDA’s software-defined moat while simultaneously overestimating how much MU’s new contract structure eliminates cyclicality. If hyperscaler capex stays elevated, NVDA can compound through mix and attach rates even if headline accelerator growth moderates. What would break the thesis: Rubin/PC ramp delays, weaker-than-expected gross margin progression, or any sign that hyperscaler procurement is shifting toward internal silicon faster than expected.

Key Decisions for Investors

• Long NVDA / short INTC or AMD as a 1-3 month relative-value expression; the setup favors the company that can monetize the whole stack versus names with direct share-loss exposure. Use a tight stop if AMD regains design-win momentum or if INTC shows credible roadmap execution.
• Add to NVDA on pullbacks rather than chasing strength; the next catalyst path is product ramp confirmation over the next 2 quarters. Risk/reward is favorable as long as management keeps gross margin and supply visibility intact.
• Do not short MU into this tape; instead, treat it as a buy-on-dip candidate over 6-12 months if memory pricing remains firm and the contract backlog converts into margin stability. Falsify if DRAM/NAND pricing starts rolling over or if guidance points to a faster normalization.
• Watch DELL and HPE as second-order winners from AI infrastructure buildout; they can benefit from unit growth and integration demand, but margins may lag NVDA due to tighter hardware economics. Prefer them only as lower-beta complements, not core longs.
• Set an alert for any guidance revision tied to Rubin volume production or AI capex commentary from hyperscalers; if either disappoints, reduce NVDA exposure and expect multiple compression across the AI hardware complex.