Intel Xeon 6 in Nvidia DGX Rubin NVL8: possibly the Xeon 6776P

Intel and Nvidia’s partnership surfaces at GTC

Last year, Intel and Nvidia spoke publicly about working together. At today’s GTC event, the collaboration has produced its first genuinely notable results: it turns out that some of Nvidia’s newer products now ship with Intel processors as part of the overall platform.

Xeon 6 in Nvidia DGX Rubin NVL8 for AI inference

More specifically, the DGX Rubin NVL8 systems are built around Xeon 6 CPUs, where the processors will handle AI inference. It seems Nvidia may have decided this route made more sense than relying on its own newer Vera CPUs.

"Artificial intelligence is shifting from large-scale training to real-time inference everywhere, driven by agentic AI and reasoning systems. In this new era, the central processor is critically important. It manages orchestration, memory access, model security and throughput in GPU-accelerated systems. Intel Xeon 6 delivers leading performance, efficiency and compatibility with the vast x86 software ecosystem that customers rely on to scale inference workloads"

The Xeon 6 capabilities Intel highlights

Intel itself points to several headline strengths of its new CPUs:

• Up to 8 TB of system memory to support large models
• MRDIMM technology, delivering three times higher memory bandwidth from one generation to the next, improving data transfer speeds to GPUs
• Industry-leading PCIe 5.0 lanes to support AI accelerators and other devices
• Confidential computing across CPU–GPU data paths using an Encrypted Bounce Buffer
• Hardware isolation to protect data and AI models while they are in use

Possibly the Xeon 6776P

Nvidia has not specified exactly which CPU variant is used in its system. It could well be a semi-custom solution. However, there is a relevant precedent for this Intel–Nvidia pairing: the Xeon 6776P is used in the GB300 Blackwell platform. It cannot be ruled out that this 64-core CPU has also been carried over into the new system.