Alvin Lang
Jun 01, 2026 04:17
NVIDIA’s Vera Rubin platform, now in full production, promises 10x agent AI throughput, targeting next-gen AI factories globally.
NVIDIA has announced that its Vera Rubin platform is ramping into full production, positioning itself as the backbone for next-generation “agentic AI” factories worldwide. The announcement, made during GTC Taipei, highlights a platform capable of delivering 10x agent throughput at scale compared to the prior-generation NVIDIA Grace Blackwell architecture. Initial shipments are set to begin this fall.
Vera Rubin represents NVIDIA’s most integrated AI infrastructure to date, combining seven cutting-edge components such as the Vera CPU, Rubin GPU, and Spectrum-6 Ethernet switches into a single rack-scale system. The NVL72 configuration, a flagship model, packs 72 GPUs into a highly optimized supercomputing rack, aimed at supporting complex reasoning and generative AI workloads with unprecedented efficiency.
“Agentic AI is a new kind of workload. One prompt can launch a thousand-step journey of reasoning, retrieval, tool use, and response generation,” said Jensen Huang, NVIDIA’s founder and CEO. “Vera Rubin was built for this moment — an AI factory engine that delivers intelligence at scale, with the performance, efficiency, and security needed to power the next industrial revolution.”
Scaling AI Factories Globally
The Vera Rubin platform is already seeing widespread adoption, with over 150 partners in Taiwan and 350 factories across 30 countries ramping up production. Industry heavyweights like Dell, HPE, Lenovo, and Supermicro are among the system builders adopting Vera Rubin to accelerate AI operations. The platform’s open-source MGX design allows for flexibility across diverse AI workloads, making it a game-changer for hyperscalers and cloud providers.
A key innovation in the platform is NVIDIA Spectrum-X Ethernet Photonics, a co-packaged-optics switch technology that boosts power efficiency by 5x and reduces AI deployment time by 1.3x compared to traditional networking solutions. This networking breakthrough enables the creation of “million-GPU factories” for large-scale AI deployments, with companies like Oracle Cloud Infrastructure and CoreWeave among early adopters.
Market and Strategic Implications
With the Vera Rubin platform, NVIDIA is doubling down on its strategy to dominate the AI ecosystem. By integrating compute, networking, and storage into a unified system, the company is transforming data centers from traditional compute clusters into fully functional AI factories. This approach addresses critical bottlenecks in power consumption, memory bandwidth, and multi-agent performance — all essential for trillion-parameter AI models and real-time reasoning tasks.
The platform’s potential is underscored by its cost efficiency. NVIDIA claims that Vera Rubin delivers up to 10x lower cost per token for AI inference compared to prior systems. This could significantly reduce operational costs for major AI labs like OpenAI, Anthropic, and Meta, which are already planning to deploy the platform for training multimodal models and serving long-context AI workloads.
Security and Performance at Scale
Security is another cornerstone of the Vera Rubin system. It incorporates NVIDIA’s Confidential Computing technology to create a trusted execution environment, encrypting data across high-speed interconnects and ensuring tamper-proof operations. This level of security is increasingly critical as AI factories handle proprietary and regulated data.
The platform also integrates the NVIDIA DOCA software layer, which enforces zero-trust policies, runtime threat detection, and end-to-end encryption at up to 800Gb/s speeds. This ensures robust multi-tenant isolation and operational reliability at scale — key for enterprises looking to roll out AI infrastructure globally.
What’s Next?
As production shipments of Vera Rubin begin this fall, NVIDIA’s market dominance in AI infrastructure appears set to deepen. The platform’s promise of unparalleled efficiency and scalability could further boost its already towering valuation, with the company’s market cap sitting at $5.15 trillion as of May 2026.
For traders and investors, NVIDIA’s strategy to control the AI stack from silicon to systems could be a long-term catalyst. As AI adoption accelerates, the Vera Rubin platform may establish itself as the gold standard for AI factories, making NVIDIA a critical player in the next wave of technological innovation.
Image source: Shutterstock
Be the first to comment