Nvidia invests $5B in Intel to co-develop AI data center CPUs and RTX-powered PC chips
Nvidia is taking a $5 billion equity stake in Intel and teaming up to build “multiple generations” of products spanning data centers and PCs—an alliance that pushed Intel shares up as much as 30% intraday and 22.8% at the close on September 18. Nvidia is purchasing stock at $23.28 per share as part of the deal.
The announcement
- Nvidia and Intel will co-develop custom x86 CPUs for Nvidia’s AI infrastructure platforms, tightly coupling with Nvidia GPUs and networking.
- For PCs, Intel will build new x86 SoCs that incorporate Nvidia RTX GPU chiplets.
- Nvidia CEO Jensen Huang called it a “historic collaboration,” estimating a large addressable market for the combined offerings.
- The investment is subject to regulatory approvals; no product release dates were announced.
Technical details
- Data center: Intel to design custom x86 CPUs that connect to Nvidia accelerators via NVLink for higher-bandwidth CPU–GPU links than PCIe, aimed at large-scale AI workloads.
- PC: Intel’s x86 “RTX SoCs” will integrate Nvidia GPU chiplets, targeting a wide range of laptops and desktops.
- Scope: Collaboration focuses on product integration; it does not commit Nvidia to manufacture on Intel Foundry at this time.
Impact and applications
- Data centers could see denser, more efficient AI racks by pairing Intel CPUs with Nvidia GPUs, potentially accelerating training and inference at hyperscalers.
- In PCs, the move pressures AMD across both CPU and GPU segments, while giving OEMs a new high-performance x86 option for AI-enabled laptops.
- Market reaction: Intel surged up to 30% intraday; Nvidia gained ~3%.
What’s next
- The companies signaled a multi-year roadmap with “multiple generations” of products; fuller specifications and timelines are expected in the coming quarters.
- Watch for ecosystem updates (software stacks, SDKs, board designs) and whether future phases broaden into packaging or foundry work.
In a year defined by AI hardware realignment, the Nvidia–Intel pact reshapes the competitive map from cloud to client—melding Nvidia’s acceleration stack with Intel’s x86 footprint and setting up a new front in the race to power AI everywhere.