It’s exhausting to say whether or not Nvidia has ever really been refined with its bulletins. At final 12 months’s CES, CEO and founder Jensen Huang surprised the business with the debut of the GeForce RTX 50 collection alongside Nvidia Cosmos, its bold world-model initiative. This 12 months’s present was extra restrained on the buyer GPU entrance, however the message to CES 2026 attendees was nonetheless unmistakable: Nvidia needs all of it.
CES 2026: Dell XPS is again from the lifeless
“All” is not hyperbole. Nvidia is now the primary firm ever to surpass a $5 trillion valuation — an nearly inconceivable determine — and Huang and firm present no indicators of slowing down. The firm’s ambitions now span factories, autonomous autos, robotics, and almost any area that may be educated, examined, or perfected in simulation earlier than ever touching the actual world. If one thing may be modeled, Nvidia needs to energy it.
Nvidia’s actual obsession is bodily AI
The greatest buzzword of the evening was “physical AI,” Nvidia’s time period for AI methods that don’t simply generate content material however really act. These fashions are educated in digital environments utilizing artificial information, then deployed into bodily machines as soon as they’ve realized how the world works.
Credit: Bridget Bennett/Bloomberg by way of Getty Images
Huang showcased Cosmos, a world basis mannequin able to simulating environments and predicting motion, alongside Alpamayo, a reasoning mannequin particularly designed for autonomous driving. This is the tech Nvidia says will energy robots, industrial automation, and self-driving autos, as demonstrated by the Mercedes-Benz CLA, which was proven working AI-defined driving on stage. The firm additionally revealed plans to check its personal robotaxi service with a associate as quickly as 2027, utilizing Level 4 autonomous autos able to driving with out human intervention in restricted areas.
Nvidia hasn’t introduced the place the service will launch or with whom it’s partnering, however the transfer alerts a shift from being a behind-the-scenes provider to actively collaborating within the self-driving race. Huang has already described robotics — together with autonomous autos — as Nvidia’s second-most vital progress class after AI itself.
Mashable Light Speed
No New GPUs
If you had been ready for brand spanking new client GPUs, you in all probability seen in a short time that there weren’t any. Nvidia didn’t announce a single new GeForce card, and that felt solely intentional. Instead, Huang spent a lot of the keynote speaking about Rubin, Nvidia’s next-generation AI platform that’s already in full manufacturing.
CES 2026: Asus’ ROG Zephyrus Duo is lastly a correct dual-screen gaming laptop computer
Rubin is described as greater than only a chip, however a complete system. GPUs, CPUs, networking, and storage, all designed collectively to deal with the immense (and environment-altering) compute calls for of contemporary AI fashions at information heart scale. Nvidia framed this as important to maintaining with skyrocketing AI demand, the place coaching prices, vitality use, and bottlenecks have gotten existential issues.
The absence of gaming {hardware} should not be thought-about a snub, however it’s clear that Nvidia is not pushed by players. It’s form of been clear that is been the case for some time, however as we speak’s convention actually drove the nail within the coffin. Instead, the corporate’s ambitions are pushed by hyperscalers, governments, and anybody making an attempt to automate all the pieces that strikes.
‘Open’ AI, powered by Nvidia {hardware}
The third main takeaway was Nvidia’s ongoing push to make itself unavoidable by openness — or not less than Nvidia’s model of it. Huang repeatedly emphasised that the corporate isn’t simply promoting {hardware}, however open AI fashions that builders can really use, fine-tune, and deploy (to not be confused with ChatGPT developer OpenAI). Nvidia now has open fashions spanning healthcare, local weather science, robotics, embodied intelligence, reasoning AI, and autonomous driving, all educated on Nvidia supercomputers and launched as foundational constructing blocks. They’ve virtually turn out to be the corn of tech.
Even private AI brokers bought some stage time, with demos of native brokers working on Nvidia’s DGX Spark {hardware}. Nvidia goals to be the platform beneath each AI system, from large information facilities to particular person desktops. It’s a sublime technique — promote openness, however nonetheless personal the pipes.
Taken collectively, the keynote felt like a declaration. Nvidia isn’t chasing CES hype cycles any extra. It’s positioning itself because the spine of an AI-powered world, the place an important bulletins don’t occur on stage, and essentially the most impactful merchandise aren’t meant to suit beneath your desk.