On June 1, 2026, at 11:00 AM, Jensen Huang, founder and CEO of NVIDIA, delivered the keynote speech at GTC Taipei 2026 in Taipei, China. He announced that the Vera Rubin architecture has fully entered the mass production stage. This architecture utilizes HBM from Micron, SK Hynix, and Samsung, providing a cabinet-level integrated AI factory foundation with significantly reduced assembly time. Huang stated that Agent AI will become a key focus for future development, marking the arrival of the era of agentic AI. Additionally, NVIDIA's Vera BlueField-4 STX architecture provides chip-level secure storage and processing capabilities for agentic AI, enhancing Token throughput, reducing data interaction latency, and offering a blueprint for high-speed data access and seamless acceleration in AI factories.
