NVIDIA CEO Jensen Huang took the stage at the GTC 2026 conference in San Jose this week, delivering a keynote packed with announcements that underscored the company's commanding position in the artificial intelligence hardware market. The centerpiece of the presentation was the unveiling of Vera Rubin, NVIDIA's next-generation GPU architecture, which is expected to ship to customers later this year. Huang described Vera Rubin as a generational leap in performance and efficiency, designed to meet the surging global demand for AI computing power.
Among the most significant reveals was the Groq 3 Language Processing Unit, the first chip to emerge from NVIDIA's $20 billion acquisition of Groq finalized in December 2025. The Groq 3 LPU is engineered for ultra-fast inference workloads, and shipments are expected to begin in the third quarter of this year. NVIDIA also introduced the Groq 3 LPX rack, a high-density system capable of housing 256 LPUs in a single enclosure, targeting hyperscale data center operators seeking maximum throughput.
Huang also showed a prototype of Kyber, NVIDIA's next major rack architecture that integrates 144 GPUs in a vertical configuration to achieve greater density and lower latency. Kyber will be available in the Vera Rubin Ultra configuration, with shipments planned for 2027. In a bold projection, Huang stated that NVIDIA expects to receive one trillion dollars in combined orders for its Blackwell and Vera Rubin chip families through 2027, reflecting the extraordinary pace of investment in AI infrastructure worldwide.
On the software side, NVIDIA announced DLSS 5, the latest iteration of its Deep Learning Super Sampling technology for gaming and professional visualization. The company also launched NemoClaw, an open-source stack for developing autonomous AI agents, paired with the DGX Spark and DGX Station hardware platforms. Huang endorsed the OpenClaw AI agent platform as a seismic shift in how enterprises will deploy intelligent automation, signaling NVIDIA's broader ambitions beyond hardware.
The automotive sector featured prominently in the keynote. NVIDIA revealed a partnership with Uber to power its fleet with NVIDIA Drive AV software across 28 cities on four continents by 2028. Additionally, automakers including Nissan, BYD, Geely, Isuzu, and Hyundai are building Level 4 autonomous vehicles on the NVIDIA Drive Hyperion platform. Huang also confirmed that NVIDIA is restarting production of its H200 AI processor for the Chinese market, where demand has been picking up despite ongoing export restrictions.
The GTC 2026 conference, running from March 16 through 19 in San Jose, has drawn thousands of developers, researchers, and industry leaders eager to see the next wave of AI hardware and software. With the Vera Rubin architecture, the Groq 3 LPU, the Kyber rack prototype, and sweeping autonomous vehicle partnerships, NVIDIA has laid out an ambitious roadmap that cements its role at the center of the global AI revolution.
Comments