2026 is shaping up to be another big year for NVIDIA. Yesterday, CEO Jensen Huang took the stage to kick off CES in Las Vegas, delivering a nearly two-hour keynote. The NVIDIA CES 2026 speech was packed with announcements on AI, robotics, and enterprise-grade silicon.
If you follow AI or gaming, this is your roadmap for what’s coming, not just from NVIDIA, but for the broader world of intelligent machines, advanced graphics, and next-gen gaming in 2026.
Cute Robots, Serious Tech
Two cute reactive robot companions joined Jensen on the stage, but they were far more than a gimmick. Powered by Cosmos AI, they’re proof that NVIDIA’s autonomous technology can work in the real world. These robots aren’t just for show: the same AI will power machines for Uber Eats, LG, Caterpillar, and Boston Dynamics. A future where intelligent, self-sufficient robots become part of daily life is closer than ever.
NVIDIA Alpamayo: AI Behind the Wheel
Next up: Jensen introduced Alpamayo, NVIDIA’s reasoning-based AI for self-driving cars. Unlike reactive systems, Alpamayo predicts hazards, plans maneuvers, and thinks ahead like a human driver. In a demo, a Mercedes-Benz CLA navigated city streets entirely on its own. It even spotted a ball rolling into the road and recognized it as a potentially dangerous situation; a child could be following behind. In this case, one was.
Thanks to its NVIDA “brains,” the car completed its route safely. Huang said the goal of Alpamayo is to bring “humanlike thinking to AV decision making.” So the system doesn’t simply react to its environment but uses it to make judgments, predictions, and change its actions.
An NVIDIA X Mercedes-Benz CLA is set to debut in the 2025. Huang has called it the “safest car in the world.”
Vera Rubin GPU: Powering AI Workloads
NVIDIA launched its new Vera Rubin GPU yesterday. And it’s not just faster; it’s a serious leap in efficiency. According to NVIDIA, Vera Rubin can handle five times the AI training workload of Blackwell, while using far fewer resources. Large “mixture of experts” (MOE) AI models that once needed a fleet of GPUs can now be trained with roughly a quarter of the hardware and a fraction of the token cost.
Originally expected later in the year, Rubin’s early debut comes on the heels of NVIDIA’s record-breaking data center revenue, up 66% year-over-year, fueled by the success of Blackwell and Blackwell Ultra. The Vera Rubin architecture is designed to meet the skyrocketing demand for AI, and products leveraging Rubin will start rolling out via NVIDIA’s partners in the second half of 2026.
Gaming and Creators: DLSS 4.5 and G-SYNC Pulsar
NVIDA didn’t forget gamers. DLSS 4.5 Super Resolution brings a second-generation transformer model and 6X Dynamic Multi Frame Generation, improving image quality and frame rates in over 400 games. Spring 2026 will see Multi Frame Generation roll out for the latest RTX 50 series GPUs.
Upcoming titles like PRAGMATA, Resident Evil Requiem, 007 First Light, and Phantom Blade Zero will leverage DLSS 4.5 and path tracing. Meanwhile, NVIDIA ACE brings smarter AI teammates and in-game systems. And for display tech, the new G-SYNC Pulsar monitors deliver over 1,000 Hz effective motion clarity. They’ll make gameplay smooth and ultra-responsive.
Looking Ahead
NVIDIA’s CES 2026 keynote made one thing clear: the company is reshaping how machines interact with the world. From autonomous vehicles thinking ahead with Alpamayo, to Vera Rubin GPUs powering massive AI workloads more efficiently, the company is setting the pace for what’s possible in 2026 and beyond.