NVIDIA founder and CEO Jensen Huang took the stage at the Fontainebleau Las Vegas to open CES 2026, declaring that AI is scaling into every domain and every device.Computing has been fundamentally reshaped as a result of accelerated computing, as a result of artificial intelligence, Huang said. What that means is some $10 trillion or so of the last decade of computing is now being modernized to this new way of doing computing.
Huang unveiled Rubin, NVIDIA's first extreme-codesigned, six-chip AI platform now in full production, and introduced Alpamayo, an open reasoning model family for autonomous vehicle development - part of a sweeping push to bring AI into every domain.
With Rubin, NVIDIA aims to push AI to the next frontier while slashing the cost of generating tokens to roughly one-tenth that of the previous platform, Huang said, making large-scale AI far more economical to deploy.
Huang also emphasized the role of NVIDIA open models across every domain, trained on NVIDIA supercomputers, forming a global ecosystem of intelligence that developers and enterprises can build on.
Every single six months, a new model is emerging, and these models are getting smarter and smarter, Huang said. Because of that, you could see the number of downloads has exploded.
Find all NVIDIA news from CES in this online press kit.
A New Engine for Intelligence: The Rubin Platform Introducing the audience to pioneering American astronomer Vera Rubin, after whom NVIDIA named its next-generation computing platform, Huang announced that the NVIDIA Rubin platform, the successor to the record breaking NVIDIA Blackwell architecture and the company's first extreme-codesigned, six chip AI platform, is now in full production.
Built from the data center outward, Rubin platform components span:
Rubin GPUs with 50 petaflops of NVFP4 inference
Vera CPUs engineered for data movement and agentic processing
NVLink 6 scale up networking
Spectrum X Ethernet Photonics scale out networking
ConnectX 9 SuperNICs
BlueField 4 DPUs
Extreme codesign - designing all these components together - is essential because scaling AI to gigascale requires tightly integrated innovation across chips, trays, racks, networking, storage and software to eliminate bottlenecks and dramatically reduce the costs of training and inference, Huang explained.
He also introduced AI-native storage with NVIDIA Inference Context Memory Storage Platform - an AI native KV cache tier that boosts long context inference with 5x higher tokens per second, 5x better performance per TCO dollar and 5x better power efficiency.
Put it all together and the Rubin platform promises to dramatically accelerate AI innovation, delivering AI tokens at one-tenth the cost. The faster you train AI models, the faster you can get the next frontier out to the world, Huang said. This is your time to market. This is technology leadership.
Open Models for All
NVIDIA's open models - trained on NVIDIA's own supercomputers - are powering breakthroughs across healthcare, climate science, robotics, embodied intelligence and autonomous driving.
Now on top of this platform, NVIDIA is a frontier AI model builder, and we build it in a very special way. We build it completely in the open so that we can enable every company, every industry, every country, to be part of this AI revolution.
The portfolio spans six domains - Clara for healthcare, Earth-2 for climate science, Nemotron for reasoning and multimodal AI, Cosmos for robotics and simulation, GR00T for embodied intelligence and Alpamayo for autonomous driving - creating a foundation for innovation across industries.
These models are open to the world, Huang said, underscoring NVIDIA's role as a frontier AI builder with world-class models topping leaderboards. You can create the model, evaluate it, guardrail it and deploy it.
AI on Every Desk: RTX, DGX Spark and Personal Agents Huang emphasized that AI's future is not only about supercomputers - it's personal.
Huang showed a demo featuring a personalized AI agent running locally on the NVIDIA DGX Spark desktop supercomputer and embodied through a Reachy Mini robot using Hugging Face models - showing how open models, model routing and local execution turn agents into responsive, physical collaborators.
The amazing thing is that is utterly trivial now, but yet, just a couple of years ago, that would have been impossible, absolutely unimaginable, Huang said.
The world's leading enterprises are integrating NVIDIA AI to power their products, Huang said, citing companies including Palantir, ServiceNow, Snowflake, CodeRabbit, CrowdStrike, NetApp and Semantec.
Whether it's Palantir or ServiceNow or Snowflake - and many other companies that we're working with - the agentic system is the interface.
At CES, NVIDIA also announced that DGX Spark delivers up to 2.6x performance for large models, with new support for Lightricks LTX 2 and FLUX image models, and upcoming NVIDIA AI Enterprise availability.
Physical AI
AI is now grounded in the physical world, through NVIDIA's technologies for training, inference and edge computing.
These systems can be trained on synthetic data in virtual worlds long before interacting with the real world.
Huang showcased NVIDIA Cosmos open world foundation models trained on videos, robotics data and simulation. Cosmos:
Generates realistic videos from a single image
Synthesizes multi camera driving scenarios
Models edge case environments from scenario prompts
Performs physical reasoning and trajectory prediction
Drives interactive, closed loop simulation
Advancing this story, Huang announced Alpamayo, an open portfolio of reasoning vision language action models, simulation blueprints and datasets enabling level 4 capable autonomy. This includes:
Alpamayo R1 - the first open
