In 2026, NVIDIA’s GTC conference opened its doors in San Jose. During the event, NVIDIA CEO Jensen Huang unveiled a bold vision: NVIDIA is evolving from a chipmaker into an “AI power plant,” dedicated to reshaping computing infrastructure for the AI age. He predicted that by 2027, the computing market will surge past $1 trillion, with NVIDIA leading the charge in delivering the world’s most cost-effective computing solution per token.

On the accelerated computing front, Huang introduced the Vera Rubin platform as the new cornerstone of computing dominance. Designed to link up to 144 GPUs, the platform promises unparalleled energy efficiency and a fivefold revenue output ratio. In graphics and ecosystems, CUDA marked its 20th anniversary, boasting hundreds of millions of GPU installations worldwide, while also unveiling DLSS 5 technology.

In the realm of agentic AI, NVIDIA rolled out NemoClaw, a dedicated operating system foundation tailored for agentic computers. Turning to physical AI, Huang underscored its significance, announcing that multiple automakers have joined NVIDIA’s autonomous driving platform. The company also released a general-purpose foundational model for humanoid robots and updated its robotics platform.

Additionally, Huang delved into CUDA’s evolution and impact, NVIDIA’s collaborations with global enterprises and cloud services, and the company’s unique blend of vertical integration and horizontal openness. He highlighted achievements from the “Year of Inference,” a decade of computing power advancements, architectural performance leaps, and the strategic acquisition of Groq to fuse cutting-edge technologies. Huang also outlined a roadmap to transcend Moore’s Law, the launch of the DSX platform, support for the OpenClaw ecosystem, and breakthroughs in physical AI and robotics.