Nvidia: Vera Rubin Production Features 5-Minute Assembly
TAIPEI (TVBS News) — Nvidia announced that its next-generation AI computing platform, Vera Rubin, has entered full production, according to remarks made by CEO Jensen Huang during Nvidia GTC Taipei at COMPUTEX 2026. "We're in volume production now," Huang said during earlier presentations introducing the platform. At GTC Taipei, Huang said Vera Rubin was designed as a rack-scale system for the era of AI agents and large-scale AI factories.
According to Nvidia, Vera Rubin integrates multiple technologies into a single platform, including the Vera CPU, Rubin GPU, ConnectX-9 SuperNIC, BlueField-4 DPU, networking technologies, storage systems, and software infrastructure. Huang said the platform was developed through close collaboration across Nvidia and its manufacturing ecosystem. He described Vera Rubin as a multi-rack system designed to support AI workloads involving reasoning, planning, memory management, and tool usage.
During the presentation, Huang highlighted the role of Taiwan's supply chain partners. He expressed his deep gratitude to local companies for supporting the development and manufacturing of AI infrastructure. According to Huang, the supply chain supporting Vera Rubin is significantly larger than that of previous generations and includes advanced semiconductor manufacturing, packaging, memory, networking, and system-integration technologies.
Nvidia also disclosed improvements in manufacturing efficiency. Huang said assembling a Grace Blackwell rack previously required about two hours. He said design optimization and supply-chain improvements have reduced assembly time for a Vera Rubin rack to approximately five minutes. Based on NVIDIA's figures, the assembly process is about 24 times faster than that of the Grace Blackwell generation.
Nvidia said Vera Rubin incorporates confidential-computing and security technologies designed to protect AI workloads, storage systems, and data processing environments. The company said the platform is intended for large-scale AI computing deployments and is now being manufactured through its global supply chain network.