Intel and NVIDIA Co-Develop Custom Xeon CPU With NVLink Integration

Posted on Edited on
Intel CEO Lip-Bu Tan reveals a custom Xeon CPU in development with full NVIDIA NVLink integration, targeting best-in-class AI server performance.

Quick Report

Intel CEO Lip-Bu Tan has confirmed that Intel and NVIDIA are co-developing a custom Xeon CPU with full NVLink integration. This collaboration aims to deliver best-in-class x86 performance for AI host nodes, directly connecting Xeon CPUs to NVIDIA GPUs for high-bandwidth, low-latency data transfer.

The announcement was made during Intel's Q4 2025 earnings call, with details still scarce. The project is expected to leverage future Xeon technologies, possibly Diamond Rapids or Coral Rapids, and could launch as early as 2027. This move signals a major shift in AI server architecture, strengthening the Intel-NVIDIA partnership in the datacenter space.

Written using GitHub Copilot GPT-4.1 in agentic mode instructed to follow current codebase style and conventions for writing articles.

Source(s)

  • TPU
  • Intel News
  • Wccftech