Marvell spikes after Nvidia invests $2 billion in the custom chip maker
And YOU get $2 billion and YOU get $2 billion.
Shares of Marvell Technology are spiking in premarket trading after Nvidia announced that it’s investing $2 billion in the custom chip company as part of another strategic partnership.
The pact will see Marvell “provide custom XPUs and NVLink Fusion-compatible scale-up networking,” a major step for Nvidia in demonstrating the willingness and the ability of its AI offerings to be deployed in concert with other companies’ products. In essence, Nvidia is ready for a world where data centers are composed of a mix of its GPUs as well as custom chips. Making sure those custom chips integrate well with its AI infrastructure platform will help maximize the dollars it receives for every gigawatt of data center capacity deployed.
And for Marvell, it’s a big vote of confidence in its custom chip and networking business.
This alliance builds on Nvidia CEO Jensen Huang’s keynote address at the GPU Technology Conference this month, where he repeatedly stressed that the chip designer is both vertically integrated (that is, offers all the solutions you need, not just GPUs) and also horizontally open (read: willing to integrate its offerings into whatever your technology stack happens to be).
Other parts of the AI ecosystem are catching a bid on this news, like data center companies Nebius, CoreWeave, Cipher Digital, Applied Digital, and IREN, as well as optical communications upstart POET Technologies.
The chip designer’s most recent strategic investment was in neocloud Nebius, also for $2 billion.
In addition, Nvidia recently invested (you guessed it!) $2 billion apiece in advanced optics companies Coherent and Lumentum as part of deals that included purchase commitments (unlike this partnership with Marvell).
However, Marvell and Nvidia will also be collaborating on silicon photonics technology, according to the press release.
