Marvell's earnings and the 8% drop
Marvell (MRVL) posted strong numbers and raised guidance, yet the stock fell about 8% after the print. The stock had climbed hard going into earnings, so the results were not enough to satisfy that run-up.
The drop traces back to the Alphabet (GOOGL) deal. Alphabet said Marvell would make some of its chips, and the market read this as mostly the network-focused chips. Investors expected more of that revenue to show up now, and more importantly in the guidance. That guidance is getting pushed out, so the payoff comes later than hoped. Nothing broke; not enough happened.
The bull case: guidance is rising, Google revenue comes later
The print, the results, and the guidance all looked good. Marvell raised its next-year top-line guide by about 10%, and the Google numbers are not even in that guide yet. The company is firing on all cylinders across data center silicon, with accelerating demand for its portfolio over the next two years. Past those two years, the Google revenue starts to ramp up and become attractive.
I put fair value at $300 a share. The market has loved MRVL this year, but there is more to go. The market is still underpricing the longer-term growth from the Google deal.
The Nvidia and Broadcom context
Jensen Huang once pointed at Marvell from stage and called it a possible next big thing. Nvidia (NVDA) sits at the center of the data center web and knows the technical road map, which should give some relief to shareholders who bought in higher.
Chasing the "next Nvidia" - meaning the 2021 Nvidia that expanded geometrically over and over - is a poor bet. Good luck with that. Owning good companies that provide good services and stay in demand is good enough, and investors buying these names should size up their expectations to that.
Custom silicon: Broadcom or Marvell?
The choice in custom high-end chips is often framed as Broadcom (AVGO) versus Marvell. Both are attractive right now, and both rank as top picks in the semiconductor and tech hardware space. XPUs and custom ASICs are a strong long-term trend inside the AI trade. Even with the large growth expected for AI infrastructure broadly, these XPUs should take share from Nvidia's generalized GPUs and grow even faster as more customers adopt them and as customers like Google adopt more chip iterations. It is an "all of the above" call: Broadcom and Marvell both show attractive growth, and the market is still underpricing the long-term opportunity for both.
Takeaway
MRVL's selloff came from timing, not weakness - the Google revenue is real but arrives beyond the next two years. Hold a longer view than the next 15 minutes, and the custom-chip story for both Marvell and Broadcom still has room.
