Marvell Technology and the custom silicon premium: Probabilistic assessment post-earnings
Following the recent drop in $MRVL to $122.28, the market seems to be re-evaluating the near-term timeline for non-GPU custom silicon ramps. While the narrative around custom ASICs for hyper-scalers remains structurally sound, the execution timeline is notoriously chunky. My base case for month-end is a consolidated range between $115 and $130, rather than an immediate V-shaped recovery to the previous highs.
I put the odds of $MRVL reclaiming the $135 level before the next monthly options expiry at roughly 25%. My reasoning is simple: the capital expenditure is there, but lead times for 3nm designs are stretching, and the broader semiconductor index is showing signs of fatigue. Conversely, there is about a 60% probability we drift sideways to build a base around the current $120.56–122 area, which represents key structural support from last quarter. If that fails, a brief wash-out to $110 is on the cards, but I expect long-term sovereign and institutional buyers would step in quickly at those multiples.
I think your $115 to $130 range is realistic, especially since the hyperscaler CapEx is notorious for lumpy deployments that the market always misjudges. My concern is whether the Broadcom comparison keeps dragging Marvell down if the next quarterly guide is anything less than perfect. Are you holding any downside protection here, or just planning to buy more if it touches the lower end of your range?