The part I like is who carries the risk. GF's Q2 capex was $411M vs $159M a year earlier, mostly into silicon photonics and SiGe, and that pushed free cash flow to about zero. A multi-year Marvell commitment is what turns that spend into booked wafers for 2027 and 2028.

Your trash bin strategy is really great.

Makes sense for GFS, which has been the forgotten foundry this year. The stock is still about half its high, so a multi-year capacity deal with a customer like Marvell is the kind of news that changes that story if more follow.