On September 18, 2025, Nvidia and Intel announced a multi-generation partnership to build new chips for data centers and PCs.
Nvidia will invest $5 billion in Intel stock at $23.28 per share. Nvidia CEO Jensen Huang said the collaboration couples Nvidia’s AI stack with Intel’s CPUs, while Intel CEO Lip-Bu Tan noted Intel’s technologies and packaging will complement Nvidia’s strengths in accelerated computing.
The agreement includes two product lines. Intel will design and manufacture Nvidia-customized x86 server CPUs for data centers. For PCs, Intel will produce x86 system-on-chips integrating Nvidia GeForce RTX GPU chiplets, branded as Intel x86 RTX SoCs. These processors will target thin-and-light gaming laptops and compact desktops that require strong CPU-GPU integration.
The new chips will pair Intel x86 cores with RTX GPU chiplets connected by NVLink, delivering up to 14 times PCIe bandwidth with lower latency. They will also support uniform memory access, allowing the CPU and GPU to share system memory. Unlike Intel’s 2017 Kaby Lake-G with AMD graphics, these chips offer faster interconnects and shared memory. Nvidia and Intel will handle their own drivers, and Intel will sell the processors.
Because the integrated GPU is GeForce RTX, future Intel PCs will support ray tracing, AI graphics, and DLSS without discrete cards. RTX Tensor Cores will provide stronger AI throughput than Intel’s NPUs, expanding CUDA and Nvidia AI software on Intel systems.
Analysts describe the deal as significant. For Nvidia, it provides entry into x86 systems. For Intel, it strengthens foundry credibility and graphics capabilities. For AMD, it raises competition across CPUs, GPUs, and consoles. Consumers could benefit from single-chip solutions delivering discrete-class performance in smaller devices.