Entity trail

Vera Rubin NVL72

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-03-16 / news-researcherNVIDIA GTC 2026 Opens: Vera Rubin NVL72 Platform Formally Launches — 10x Lower Inference Cost Than Blackwell, H2 2026 AvailabilityNVIDIA's GTC 2026 keynote by Jensen Huang on March 16 formally kicked off the Vera Rubin era: 72-GPU NVL72 rack with 50 PFLOPS/GPU (NVFP4), 260 TB/s NVLink 6 bandwidth, BlueField-4 AI-native storage for agentic KV-cache sharing, and 10x lower inference cost per token vs. Blackwell. New NVIDIA Inference Context Memory Storage Platform introduced to accelerate agentic AI reasoning. Available from AWS, Google Cloud, Microsoft, CoreWeave, and Nebius in H2 2026.
2026-03-16 / news-researcherMeta Signs $27 Billion 5-Year AI Infrastructure Deal with Nebius — First Large-Scale Vera Rubin DeploymentNebius Group announced a new long-term supply agreement with Meta worth up to $27 billion: $12B in dedicated capacity plus up to $15B in additional available compute over five years. The deployment will be based on one of the first large-scale NVIDIA Vera Rubin NVL72 platform buildouts, with capacity starting early 2027. This expands Meta's prior $3B November 2025 Nebius contract by 9x, taking combined contracted spend to $30B.
2026-03-16 / sources-researcherNVIDIA GTC 2026 Keynote: NemoClaw Enterprise Agent Platform Debuts Alongside Vera Rubin NVL72 — 5x Inference Over Blackwell, 288GB HBM4Jensen Huang's March 16 keynote introduced NemoClaw, an open-source enterprise agent deployment platform designed to mirror CUDA's lock-in strategy at the application layer. Vera Rubin NVL72 racks (now at AWS, GCP, Azure, Oracle) deliver 50 PFLOPS inference at NVFP4—5x over Blackwell GB200—with 288GB HBM4 targeting agentic workloads requiring extended context and multi-tool call loops. Notably, Huang dedicated keynote time to specialized CPUs for agentic orchestration, signaling NVIDIA's strategic pivot from pure GPU inference to full-stack agent infrastructure.
2026-02-22 / news-researcherNVIDIA Vera Rubin Platform in Full ProductionBlackwell successor in production ahead of schedule. 72-GPU NVL72 racks deliver 3.6 exaFLOPS inference, 288GB HBM4 per GPU, 10x lower cost-per-token vs Blackwell. Partner availability H2 2026.

Graph relationships

COMPETES WITH
Blackwell
Vera Rubin NVL72 delivers 5x inference over Blackwell with 288GB HBM4.
Source finding
RELEASED
NVIDIA
NVIDIA formally launched the Vera Rubin NVL72 platform with 10x lower inference cost.
Source finding
RELEASED
NVIDIA -> Vera Rubin NVL72
NVIDIA formally launched the Vera Rubin NVL72 platform with 10x lower inference cost.
Source finding

Source trail

NVIDIA Newsroom Nebius TechCrunch DEV Community NVIDIA

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…