NVIDIA Announces Partnership With Intel, NVLink & RTX To Be Integrated Into Intel Chips

Surprise – NVIDIA and Intel has jointly announced a collaboration to develop “multiple generations of custom data center and PC products,” and in particular, two of NVIDIA’s key IPs will be involved: NVLink, and GeForce RTX. Looks like we’re in for interesting times, but let’s first explain how the deal will happen.

NVIDIA & Intel Join Forces

NVIDIA says the partnership and Intel will involve through NVLink protocol on the datacenter side by combining the AI giant’s know-how with Intel’s CPU technologies and x86 ecosystem. As such, Intel will produce NVIDIA-custom x86 CPUs that will be integrated into NVIDIA’s AI infrastructure platforms and offered to the market. This likely have huge implications, especially given that NVIDIA has somewhat committed itself into Arm-based designs with recent chip releases like Grace CPU, along with the upcoming N1/N1x series destined for consumer PCs.

Speaking of consumer segments, there’s an even bigger surprise: Intel will manufacture and sell x86-based system-on-chips (SoCs) that incorporate RTX GPU chiplets, similar to AMD’s Ryzen AI Max series APUs that combine the best of Ryzen and Radeon into a single chip. There are monetary transactions involved too, as NVIDIA will invest $5 billion in Intel’s common stock at a purchase price of $23.28 per share, meaning the GPU chipmaker officially owns 4.9% of the company’s stake.

Despite this partnership, Intel’s foundry outlook continues to look bleak, as the partnership doesn’t address the foundry side of things – NVIDIA is merely involved in the future generations of chip designs from the Chipzilla, and as far as the GPU maker’s AI chips are concerned, they’re fully committed to TSMC for now. On the consumer level, there are also concerns whether the integration of RTX graphics may eventually cannibalize Intel’s Arc GPU department, which still have no meaningful footing in the discrete GPU market today.

To that end, Intel clarifies to PCWorld: “We’re not discussing specific roadmaps at this time, but the collaboration is complementary to Intel’s roadmap and Intel will continue to have GPU product offerings.” While the statement didn’t directly address Arc GPUs, at least it seems like Arc will survive in the short-term – although it’s hard to say what happens beyond Druid (“Xe4”), which is the furthest generation Intel has publicly announced in their roadmap.

Pokdepinion: This might just be a pivotal moment in the history of chipmaking – although I am worried about the future of Arc GPUs, too.