NVIDIA has announced the launch of the NVIDIA Rubin platform, “comprising six new chips designed to deliver one incredible AI supercomputer.”

According to the company, the Rubin platform – named for American astronomer Vera Rubin – uses extreme codesign across the following six chips:

• NVIDIA Vera CPU
• Rubin GPU
• NVLink 6 Switch
• ConnectX-9 SuperNIC
• BlueField-4 DPU
• Spectrum-6 Ethernet Switch

The Rubin platform also includes the following technology updates:

• Sixth-Generation NVIDIA NVLink: Delivers fast, seamless GPU-to-GPU communication.
• NVIDIA Vera CPU: Designed for agentic reasoning, NVIDIA Vera is the most power-efficient CPU for large-scale AI factories.
• NVIDIA Rubin GPU: Features a third-generation Transformer Engine with hardware-accelerated adaptive compression to deliver up to 50 petaflops of NVFP4 compute for AI inference.
• Third-Generation NVIDIA Confidential Computing: Maintains data security across CPU, GPU, and NVLink domains.
• Second-Generation RAS Engine: The Rubin platform features real-time health checks, fault tolerance, and proactive maintenance to maximize system productivity.

NVIDIA notes that the first cloud providers to use Vera Rubin-based instances in 2026 will include AWS, Google Cloud, Microsoft, and OCI. Additionally, Red Hat has announced an expanded collaboration to deliver an optimized stack for the Rubin platform, including Red Hat Enterprise Linux, Red Hat OpenShift, and Red Hat AI.

Read more at NVIDIA.