$MRVL

Does he own it?

Multiple first-person positional statements: he says outright he has positions in Marvell, lists it among "23 longs I hit this year," attributes personal returns to it ("MRVL +54% was an underperformer for me"), and notes swing-trading it on a false-news dip. No explicit % weight ever given (only SIVE got a ~0.5-1% disclosure). Treated as a safer mega-cap satellite long, not a concentrated bet.

His thesis

MRVL is a "does-everything" mini-Broadcom custom-ASIC and optical play. Core catalysts: 2-3x revenue ramp from Microsoft Maia ASICs into 2027-2029, plus the Celestial AI acquisition giving it a co-packaged-optics scale-up vector ($500M ARR 2028, $1B 2029). Repeatedly validated by NVIDIA's $2B investment, T1 hyperscaler design-ins (Amazon, Google talks), and Jensen's "$1T company" call. Crucially, MRVL is also the hub of his real obsession — mapping its Celestial light source upstream to his concentrated pick SIVE.

How his view evolved

Started as a buy-rated, undervalued post-selloff swing trade on false hyperscaler-loss news. Grew into a held core long as Celestial/NVIDIA/Maia catalysts compounded; by mid-year it had rallied ~54%+ and he confirmed holding it, framing it as a safe large-cap relative to his small-cap bets.

Key points

• Bull case: Maia ASIC ramp + Celestial CPO growth vector, endorsed by NVIDIA stake and Jensen's $1T call
• Mostly functions as the supply-chain hub linking to his true high-conviction pick SIVE (the upstream laser source)
• Held as a safer mega-cap satellite long; never disclosed a portfolio weight, lower conviction than his small-caps

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