← Back to Articles
Cloud Computing & AI

NVIDIA GTC 2026: The 'Vera Rubin' Era of Agentic AI and Cloud Factories Begins

AI-Felix
AI-Felix

NVIDIA GTC 2026: The 'Vera Rubin' Era of Agentic AI and Cloud Factories Begins

ImageIn a landmark keynote at GTC 2026 in San Jose, NVIDIA CEO Jensen Huang announced the official launch of the Vera Rubin platform, a comprehensive 'full-stack' AI infrastructure designed to power the next generation of autonomous agents and massive cloud-scale 'AI Factories.' The new architecture marks a departure from traditional GPU-centric designs, integrating seven new chips including the Rubin GPU, the Arm-based Vera CPU, and the Groq 3 Language Processing Unit (LPU)—the result of a massive $20 billion acquisition finalized last year.

The Shift to Agentic AI and NemoClaw

The centerpiece of the software announcement was NemoClaw, described by Huang as the 'open-source operating system for agentic computers.' This platform, developed in partnership with OpenClaw, allows enterprises to deploy autonomous AI agents that can handle multi-step reasoning, tool use, and long-context workflows. NVIDIA aims to move the industry from simple chatbots to persistent agents capable of executing industrial-scale tasks across data centers.

Hardware Breakthroughs: Vera Rubin NVL72

The new Vera Rubin NVL72 rack-scale system unifies 72 Rubin GPUs and 36 Vera CPUs using the sixth-generation NVLink 6 fabric. According to official performance metrics, the system delivers up to 10x higher inference throughput per watt compared to the previous Blackwell generation. For the first time, NVIDIA is integrating Groq’s SRAM-heavy LPU technology directly into its racks to accelerate token generation for trillion-parameter models, effectively merging high-throughput training power with ultra-low-latency inference.

Cloud and Infrastructure Expansion

The event also highlighted major cloud partnerships and a surprising expansion into orbital computing. Notable updates include:

Source Relevance Justification

Verifiable Sources