$POET
poet technologies3/10Does he own it?
One early tweet discloses an opportunistic swing buy on a sharp drop ("ended up buying at this level"), but he repeatedly and explicitly states he is "not personally long" POET, omits it from his detailed portfolio-weight breakdown, and calls it too risky to size into. No held-position, adding, or weight disclosure. Treated as coverage, not a conviction holding.
His thesis
POET makes optical interposers/packaging (Starlight) that backdoor into Marvell's Celestial CPO program and hyperscalers. He sees a cushion from a large cash pile after dilution, but views the Marvell-only revenue (material only ~2028) as too far out and single-customer concentration as the key risk. His core view: POET merely packages the lasers, and the real value/chokepoint sits upstream with the laser supplier Sivers, which he says POET is wildly overvalued against.
How his view evolved
Started cautiously constructive on the post-offering dip (an opportunistic buy) while flagging it as early. Grew steadily more skeptical—overvalued vs the laser maker, dilution-driven, single-customer risk—culminating in Marvell cancelling orders (-46%), validating his designed-out thesis.
Key points
- Used mainly as a node in his Sivers (SIVE) laser thesis—POET buys the lasers, real value is upstream
- Repeatedly says not personally long; calls POET overvalued, dilution-heavy, single-customer (Marvell) risk
- Marvell cancelled POET orders, crashing it ~46% and confirming his designed-out skepticism
Derived from a holistic read of his public posts — paraphrased, never quoted. See his actual posts about $POET on X ↗