NVIDIA's CEO gave us the specifics of the Intel deal during a webcast with Intel's CEO Tan, and interestingly, he discussed the prospect of teaming up with the IFS.
NVIDIA CEO Reiterates That There Won't Be Any Changes With The ARM Roadmap; TSMC Likely To Be The Primary Foundry
NVIDIA-Intel entered into a 'blockbuster' deal yesterday. The key focus was collaboration in the datacenter and consumer CPU segments, and Team Green invested $5 billion in Intel's common stock to 'celebrate' the partnership. After the deal was announced, the CEOs of both companies went live on a 'special call' taking questions from reporters of multiple media outlets. The main points discussed were about why NVIDIA entered into an x86 partnership, and whether there's a possibility of NVIDIA getting its chip needs from the Intel Foundry.
Jensen Huang disclosed that with current rack-scale products, they are able to go up to NVL72 configurations with ARM-based CPUs, while with x86 datacenter CPUs, such as Intel's Xeon, it is limited to NVL8, due to the use of a PCIe-based interface. The deal with Intel is focused towards bridging this gap, by bringing NVLink into Intel's datacenter CPUs, and this will ultimately allow NVIDIA to offer both ARM and x86-based options to its customers. NVIDIA claims that this is a $30 billion untapped market, which they will address through the partnership.
And so this architecture, the NVLink 72 rack scale architecture is only available for the Vera CPU that we build, the ARM CPU that we build. And for the x86 ecosystem, it's really unavailable except with server CPUs over PCI Express.
And that has limitations in how far you could scale these scale-up systems. And so the first opportunity is that we can now with Intel x86 CPU, integrate it directly into NVLink ecosystem and create these rack-scale AI supercomputers.
The next major point to discuss here is that NVIDIA says it will be the primary consumer of Intel's datacenter CPUs and the seller of its RTX GPU chiplets for Intel, which will be used for a PC SoC. So, the partnership isn't one-sided at all, and this means that both firms are committed to deep cooperation. Jensen was asked about why NVIDIA doesn't consider using Intel Foundry for the new chips planned, and he claimed that the firm has been working with the IFS, but, TSMC's importance cannot be ignored, which does imply that there's little room for IFS's node integration for now.
I think Lip-Bu and I would both say that TSMC is a world-class foundry. And in fact, we're both very successful customers of TSMCs. And I can't -- you just can't overstate the magic that is TSMC. But today, our conversation today, our partnership today is completely focused.
One way Intel Foundry might be involved is by collaborating on its advanced packaging technology. Jensen mentioned Intel's 'Foveros' packaging, which might be used for the PC chip to combine the RTX GPU with Intel's CPU chiplets. The plans for how NVIDIA-Intel deal will formulate aren't concrete for now, but it seems like the IFS would be involved, not just in the process technology front. Of course, this could change judging by how 18A and 14A pans out, but for now, it seems like both NVIDIA and Intel would be comfortable using TSMC.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
