SHELFCRITTER AI Digest

INFRASTRUCTURE

Rubin

infrastructuretopic-notenvidia-platform

Overview

Rubin is NVIDIA’s next-generation AI computing platform, following Blackwell, announced for broad cloud distribution starting H2 2026. The platform integrates six chips — Vera CPU, Rubin GPU, NVLink 6 switch, ConnectX-9 SuperNIC, BlueField-4 DPU, and Spectrum-6 ethernet switch — positioned as the primary infrastructure for large-scale AI training and inference deployments across major clouds and neoclouds.

Timeline

  • 2026-05-05-AI-Digest — NVIDIA formally opened the Rubin platform — six new chips spanning Vera CPU, Rubin GPU, NVLink 6 switch, ConnectX-9 SuperNIC, BlueField-4 DPU, Spectrum-6 ethernet switch — for distribution starting H2 2026 across AWS, Google Cloud, Microsoft Azure, Oracle Cloud, plus the neocloud tier (CoreWeave, Lambda, Nebius, Nscale). Headline performance claims versus Blackwell: 3.5× training throughput, 5× inference throughput, 8× power efficiency. Microsoft’s Fairwater data centre sites in Wisconsin and Atlanta reported as already operating Vera Rubin NVL72 racks. Distribution piece closed; first GA price point remains open.

Key Developments

  1. Six-Chip Integrated Platform: Unlike Blackwell’s primary GPU-centric design, Rubin bundles compute, switching, networking, and DPU infrastructure as a unified system, reducing operational complexity for hyperscalers.

  2. 3.5×/5×/8× Performance Claims: Training throughput, inference throughput, and power efficiency gains versus Blackwell establish clear generational improvement narrative.

  3. Immediate Production Deployment: Microsoft’s Fairwater sites already running Vera Rubin NVL72 racks de-risks cloud-provider adoption timelines and signals high customer confidence.

  4. Broad Cloud Distribution: Confirmed partnerships across all major clouds (AWS, GCP, Azure, OCI) plus neocloud tier (CoreWeave, Lambda, Nebius, Nscale) establish Rubin as the industry standard for H2 2026 new AI workload deployments.

  5. Deferred Pricing Disclosure: Despite confirmed distribution partnerships, first customer-facing price points remain unpublished, keeping downside customer-acquisition risk open through H2 2026.

Market Position

Rubin’s distribution announcement closes a critical loop in the Blackwell → Rubin generational shift narrative tracked since March 2026. The platform’s integration of switching, networking, and DPU alongside compute reflects NVIDIA’s effort to capture the full inference-infrastructure stack rather than just GPUs, while the immediate Microsoft production deployment suggests cloud providers are sufficiently confident in the roadmap to commit capacity before final pricing.

See Also