Intel and Nvidia’s Strategic Partnership: A New Era of Semi-Custom Solutions

Besides singing praises of their upcoming 14A process technology, Intel unveiled details on its newly inked deal with Nvidia. Under this deal, Intel will develop a specialized Xeon server processor for Nvidia, leveraging the NVLink Fusion interconnect system. Photo: WCCF Tech This is doubly interesting in light of recent leaks suggesting that Intel plans to create a special processor for portable gaming consoles. It seems the company has decided to delve deeper into semi-custom solutions akin to AMD’s strategy.

Intel’s announcement to create a processor with an integrated Nvidia graphics core further adds intrigue. The iGPU will be designed as a separate chiplet, integrated into Intel’s processors. Previously, Intel had Kaby Lake G processors with an AMD iGPU, and now we may witness something new. However, the collaboration between the two companies will be unconventional, and details are not yet clear.

Intel states: “When we look at the client side, I believe we certainly have an opportunity to create a truly new class of PC components, which excites us greatly. These relationships will be structured as follows: Nvidia provides the graphics tile on a warranty basis, so the client essentially pays them for the graphics tile, but we will be responsible for integrating that graphics tile with our CPU and bringing it to market.”

This means Intel and Nvidia are not merely creating some processors or a line of processors together. Instead, each time, the end client in the form of a laptop manufacturer will evidently have to pre-purchase a certain number of iGPUs from Nvidia and, accordingly, order an equal number of processors from Intel. Why the scheme must be this way is unclear.

Intel concluded by announcing their intention to increase prices on older Alder Lake and Raptor Lake CPUs (which have already risen) while reducing prices on Arrow Lake and Lunar Lake. This aligns with current trends where older models experience price increases, allowing space for aggressively priced new technologies. Furthermore, Nvidia has recently revealed the NVLink-C2C technology, which greatly enhances data transfer rates between chips, potentially benefiting these upcoming server processors.