"Nvidia's investment in Marvell Technology and the integration into NVLink Fusion creates a partnership that ensures revenue from custom AI accelerators designed for hyperscalers like Amazon and Google."
"The partnership mandates that every NVLink Fusion platform must include at least one Nvidia product, ensuring that custom chips designed by Marvell generate revenue for Nvidia."
"The collaboration will focus on silicon photonics, using light instead of copper to move data between chips, which is essential for the speeds required by next-generation AI clusters."
"Jensen Huang stated, 'The inference inflection has arrived,' highlighting the surge in token generation demand and the race to build AI factories."
Nvidia invested $2 billion in Marvell Technology, integrating it into the NVLink Fusion ecosystem. This partnership focuses on custom AI accelerators, silicon photonics, and 5G/6G infrastructure. Marvell will supply custom XPUs and networking, while Nvidia provides CPUs, network interface cards, and other components. The deal mandates that every custom chip designed by Marvell for hyperscalers generates revenue for Nvidia. The collaboration also emphasizes silicon photonics to meet the demands of next-generation AI clusters, with Nvidia controlling access to NVLink IP licenses.
Read at TNW | Corporates-Innovation
Unable to calculate read time
Collection
[
|
...
]