Intel and Nvidia announced a partnership last month with far-reaching implicationsbut the deal itself is rather vague. Neither Intel nor Nvidia has delved too deep into the details of this partnership, but we do know it’s going to impact the AI industry and personal computingwith Intel designing and manufacturing custom data center and client CPUs with Nvidia NVLink. That deal came with a $5 billion investment from Nvidia in Intel stock.
This deal came as a surprise initially, as Nvidia’s previous partnerships have involved Arm-based CPUs instead of x86-based processors. It also puts Intel’s budding Arc GPU platform in an interesting place. While this deal could put a stop to Intel’s discrete GPU business, it might also be the thing that truly puts Intel Arc on the map.
Intel and Nvidia: What we know
Table of Contents
The hard details on the AI deal
Intel and Nvidia have agreed to a deal where Intel will manufacture data center and client CPUs for Nvidia using the NVLink GPU interconnect.
On its face, this deal doesn’t seem to impact gaming at all and is mostly focused on both companies’ AI and data center businesses since the NVLink interconnect will allow a better connection between Intel CPUs and Nvidia GPUs in a data center rack. However, the deal does further tie Intel and Nvidia together, as Intel will also be manufacturing client SoCs that integrate Nvidia’s RTX GPU chiplets.
What purpose these client CPUs will serve, we don’t quite know just yet. The initial focus of the deal appears to be AI, but that doesn’t mean it’s the only avenue for Intel and Nvidia to take.
Could the Arc GPU team be the unlikely beneficiary?
Intel’s GPU team could use Nvidia’s gaming expertise
Intel’s nascent graphics team has been releasing discrete GPUs and integrated GPUs under the Intel Arc label for just a few years now, since the Intel Arc A750 and Arc A770 released in 2022. The Arc platform has been running iGPUs for Intel since the Meteor Lake Intel Core Ultra 100 generation, which means we’re only about 2 full generations into the Arc platform, with the third generation expected to launch early next year alongside Intel’s Panther Lake CPUs.
The deal with Nvidia allows Intel to start manufacturing SoCs with Nvidia RTX GPU cores. While these CPUs aren’t expected to land in 2026, it could mean that the Nvidia GPU cores will replace the Intel Xe GPU cores in SoCs starting as soon as 2027. Which would be bad for the Arc platform as a whole, since those Xe GPU cores power Intel’s integrated and discrete graphics platforms. This would essentially kill the Arc business.
Alternatively, the RTX GPU chiplets could work alongside Intel’s existing Xe GPU cores on these chips. Intel is reportedly looking at a dual-GPU solution for the upcoming Nova Lake architecture as is. So it’s not like a combination of Xe and RTX GPU cores on a single system-on-a-chip is out of the realm of possibility.
While Intel’s Arc platform has gotten better with each generation, to the point where the Arc B580 is a fantastic budget gaming GPU and the Arc 140V integrated graphics tile is the most powerful handheld GPU. So seeing the Arc business die off would be an utter shame. However, combining Nvidia’s RTX and DLSS technology with Intel’s fledgling Arc and XeSS systems could allow Intel’s Arc business to be a proper gaming powerhouse. Right now, other than a lack of raw power, Arc’s biggest flaw is its small library of optimized games. And Nvidia’s tech and industry connections could help bridge that gap.
AI means better gaming
No, seriously
While AI-powered graphics technology has a bad rap, upscaling and frame generation technologies like DLSS and XeSS are only getting better. And a lot of that is driven by the AI hype cycle. AI workloads need powerful GPUs after all, the same way gaming does.
So even if this deal is just about data center and client AI systems, it could have some side benefits for gaming, regardless. That was certainly true of Intel’s Lunar Lake platform. Designed to be a Copilot+ AI PC system, Intel’s Core Ultra 200V series was under the hood of the most powerful gaming handheld. While Lunar Lake’s integrated GPU was designed to efficiently handle the needs of LLM-based AI, it also offered faster graphics performance across thin and light laptops and handhelds.
So even if these Intel and Nvidia client SoCs are designed for AI workloads, they could make for solid iGPU gaming as well.
The deal could just be about AI
Or even just domestic silicon manufacturing
We are still in the early days of this partnership, and the details haven’t been fully released to the public. Intel and Nvidia may be working together solely because Intel’s Arizona and Oregon fabrication sites are ramping up domestic silicon production on the 18A (2 nanometer) process, and Nvidia is looking for a domestic manufacturing partner to avoid US tariffs. It could also be a partnership devoted entirely to cloud data center and local on-device AI, and avoid the gaming and workstation aspects of both Intel and Nvidia’s businesses entirely.
But even if this deal involves the GeForce and Arc platforms, this partnership between two tech giants could go several ways. This rundown is just speculation based on what information is publicly available. The Intel and Nvidia deal could spell the end of Intel’s Arc platform, it could reconstitute Arc into a solely workstation-focused business, or it could help improve Intel’s discrete GPU technology. We have no way of knowing just yet.
