$MU

Does he own it?

Strong direct evidence. In a Feb 7 portfolio-weighting breakdown he assigns 10% to MU within a 35% memory-supercycle bucket. Repeatedly frames himself as long the memory/AI trade in first person, says he is staying extremely long, and lists MU among names he personally owns. Explicit weight disclosure makes ownership near-certain, not mere coverage.

His thesis

MU is a centerpiece of his "memory supercycle" thesis: structural AI-driven DRAM/HBM/NAND demand, repeated price hikes lifting gross margins (he modeled ~73-75%, confirmed near 75%), and forward P/E he viewed as far too cheap. He stresses MU's "Made in America" moat versus Samsung/SK Hynix tariff risk and White House backing. Calls it the next Nvidia, expecting re-rating toward a trillion-plus valuation; treats memory demand as permanent though prices peak around 2028.

How his view evolved

Began Jan as a deeply undervalued re-rating call, reinforced through NAND/DRAM hike news and earnings that confirmed his margin estimates. Stayed bullish through Iran/macro selloffs he dismissed as noise, then declared victory as MU crossed a trillion-dollar cap by late May.

Key points

• Disclosed explicit 10% portfolio weight to MU in his memory bucket
• Thesis: structural AI memory demand plus price hikes driving margins and re-rating
• Emphasizes US-made moat and tariff advantage over Korean rivals

Derived from a holistic read of his public posts — paraphrased, never quoted. See his actual posts about $MU on X ↗