At its annual GPU Technology Conference (GTC) 2026 in San Jose, California, Jensen Huang, founder and CEO of NVIDIA, took the stage and made it clear from the first minute that the company sees no ceiling on AI’s growth.
The conference, held from March 16–19, zeroed in on practical use of AI — inference, AI factories, physical AI, and agent-based systems. NVIDIA highlighted a shift in artificial intelligence from building models to deploying them at scale across industries. “This conference is going to cover every single layer of the five-layer cake of artificial intelligence,” Huang said.
Marking 20 years of CUDA – the software platform that enables developers to use GPUs for advanced computing – Huang described it as a key driver of accelerated computing and a platform that supports the entire AI lifecycle. Huang pointed to NVIDIA’s presence across industries including automotive, healthcare, finance, robotics and telecom. “All of these different vectors of AI have platforms that NVIDIA provides,” he said, highlighting the company’s CUDA-based software ecosystem.
The #NVIDIAGTC 2026 keynote is officially LIVE! 🟢
Jensen Huang is on stage to unveil the latest breakthroughs in AI and accelerated computing. See how agentic AI, AI factories, and physical AI are powering the next generation of intelligent systems.
Watch now ⬇️…
— NVIDIA DRIVE (@NVIDIADRIVE) March 16, 2026
Huang highlighted the rise of “AI-native” companies such as OpenAI and Anthropic, as $150 billion in venture investment flowed into the sector. He added that the computing demand for NVIDIA GPUs is “off the charts,” and said “I believe computing demand has increased by 1 million times over the last few years.” The company expects AI data center and chip infrastructure to represent a $1 trillion revenue opportunity through 2027, reflecting the scale of expected investment.
A central focus this year was inference, where AI systems are deployed in real-world applications. NVIDIA said the industry is shifting from training models to running them continuously in services such as enterprise software and automation. And to support this shift, the company introduced new AI systems and said it is working with Groq to improve inference performance, widening its approach beyond its own chips.
NVIDIA said it is moving beyond selling individual chips to offering complete AI systems combining computing, networking, and software, aimed at faster deployment for businesses. Huang described the Vera Rubin platform as a fully integrated system, built to operate as one optimized AI stack. Huang announced new AI factory tools, including the Vera Rubin DSX design and Omniverse DSX Blueprint, to help companies plan and scale AI infrastructure.
NVIDIA also announced its latest accelerated computing platforms are unlocking a new era of space innovation, bringing AI compute to orbital data centers, geospatial intelligence and autonomous space operations. “AI processing across space and ground systems enables real-time sensing, decision-making and autonomy, transforming orbital data centers into instruments of discovery and spacecraft into self-navigating systems,” Huang said.
On AI agents, Huang described OpenClaw — an open-source platform that lets developers quickly build and deploy autonomous AI assistants – an open-source platform for building and deploying AI agents, adding that companies will need strategies to adopt such systems.
To help companies deploy these agents safely, NVIDIA introduced NemoClaw — a security and policy stack Huang described as capable of being “the policy engine of all the SaaS companies in the world.” NVIDIA also expanded its open model ecosystem with the launch of the Nemotron Coalition, bringing partners together around new frontier AI models.
The company highlighted growing interest in AI systems that can perform tasks with limited human input, reflecting a broader shift toward more autonomous software. Beyond software, NVIDIA pointed to expansion into robotics and industrial systems, including partnerships with robotics companies including ABB, Universal Robots and KUKA to support humanoid and industrial machines.
On the second day of the conference, March 17, the focus shifted to industry use cases, with sessions highlighting applications of AI in areas such as energy, climate research and media. NVIDIA’s programming showed how AI is being applied across sectors, reinforcing its move from development to real-world deployment
GTC 2026 reflects a wider industry transition, as AI moves from experimentation to essential infrastructure and real-world deployment.
The event featured more than 1,000 sessions and over 2,000 speakers, reflecting the scale of global interest in AI.
Also Read: NVIDIA Financial Results Hit Record Quarterly Data Center Revenue, Up 75% YoY
