Intel has reassured that it will continue to have GPU products in the future despite the new NVIDIA deal, which was announced today.
Intel Will Continue To Have Its Own GPU Products In The Future, But NVIDIA GPUs Open Room For Interesting PC Platforms
Today, NVIDIA & Intel announced a collaboration where both companies will be developing next-gen x86 chips that combine Intel's CPU prowess with NVIDIA's strong GPUs, such as RTX for client PCs and AI/HPC designs for enterprise segments.
However, during the conference between NVIDIA CEO Jensen Huang and Intel CEO Lip-Bu Tan, the blue team didn't reveal if their own GPU family would be affected by this deal. Intel currently has two major GPU lineups, the Gaudi and "Shores" family for enterprise AI and the Arc series for client/business. These two GPU families have had a hard start to take off, but have become the value king in their own segments.
If Intel is to procure future GPU chiplets from NVIDIA for its enterprise and client products, then that leaves little room for Intel's own graphics division. Intel has now come forward with its statement regarding this, with PCWorld & HotHardware receiving the following from an Intel spokesperson:
"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,"
Intel Spokesperson
Well, it looks like Intel has made it clear that, despite the new deal with NVIDIA, their own GPU product offerings will continue, and the roadmap should remain unchanged. Currently, Intel is going to launch its 3rd Gen Xe architecture, Xe3 codenamed Celestial, with Panther Lake CPUs. Panther Lake will launch soon, so it is unlikely we will see any NVIDIA DNA in those chips. However, next year's Nova Lake CPUs might be the first to leverage NVIDIA's GPUs.
If you remember, a few months back, we reported on Intel's Nova Lake-AX chip. This would be a direct competitor to halo-class SoCs like the AMD Ryzen AI MAX series. The Intel Nova Lake-AX SoC is said to feature a bigger iGPU, more cache, and some exciting innovations on the packaging level. While current rumors point towards up to 48 Xe3 iGPU cores, it looks like Intel might go back to the drawing board now that they have NVIDIA onboard and use an RTX GPU, probably based on Blackwell or a future iteration of NVIDIA client GPUs.
This would be very similar to Intel's previous attempt at an SoC, the Kaby Lake-G, which was a multi-chiplet design housing three chiplets: the Compute chiplet, the IO chiplet, and an AMD Radeon GPU chiplet based on the RX Vega architecture.
Kaby Lake-G didn't turn out to be a massive success and was plaqued with various driver-rollout issues, but it paved the way for future SoCs such as AMD's Strix Halo and Intel's AX lineup, which combine several chiplets and package them together more efficiently thanks to modern-day packaging technologies from TSMC. Intel has its own advanced packaging solutions with Foveros interconnect being stated as the one that will be used for designing these new Intel x NVIDIA chips.
While Halo chips with NVIDIA's RTX GPUs will be decent, it looks like the standard lineup will continue to use Intel's own Xe architectures, such as Xe3 and Xe4. But the main question remains: what happens to Arc discrete?
Arc discrete is something that Intel will also continue to work on its own. Using NVIDIA's GPUs for Arc discrete graphics won't make sense since they are the direct competition, and not having a discrete GPU family would just give NVIDIA an empty battlefield in the discrete segment. Their market share has climbed to 94%.
Currently, there are rumors that Intel might launch a bigger Battlemage GPU, the BMG-G31, for desktops and laptops but no conclusive details are available, nor has Intel announced such a product. Though they were also very tight-lipped about the Arc B-Series discrete offerings and only announced them a few weeks prior to launch.
Jensen also talked about how NVLINK and other technologies will be used on the AI side, though no details regarding how that would affect the Jaguar Shores AI accelerator family were provided. While the NVIDIA deal is great for Intel, we also look forward to seeing actual products being announced, which will be around CES or the next major 2026 event, so it's interesting times ahead, but even better to hear that Intel's GPU endevaors are going to continue as usual.
Follow Wccftech on Google to get more of our news coverage in your feeds.
