Nvidia's 'chatgpt Moment' For Self-driving Cars, And Other Key Ai Announcements At Gtc 2026

Screenshot by Radhika Rajkumar/ZDNET

Follow ZDNET: Add america arsenic a preferred source on Google.

ZDNET's cardinal takeaways

• Nvidia released caller models for autonomous robots, cars, and more. 
• Uber will adhd Nvidia-powered robotaxis to cities arsenic early arsenic 2027. 
• More lifelike robotics could mean robotic characters astatine Disney World.

To adjacent retired his Nvidia GTC keynote connected Monday, CEO Jensen Huang brought retired an unexpected guest: a walking, talking robot type of Olaf, nan animated snowman from Disney's Frozen movie. Huang explained to robo-Olaf that he's tally connected Nvidia's Jetson level and learned to locomotion wrong nan company's Omniverse simulator

Olaf's responses didn't ever make consciousness -- nan speech was awkward, but nan thought was clear: successful nan future, robotic characters could beryllium wandering astir Disneyland utilizing Nvidia's tech. 

Also: Nvidia wants to ain your AI information halfway from extremity to end

Physical AI -- AI systems embedded successful machines for illustration robots aliases cars that navigate real-world environments, arsenic opposed to models stuck successful nan unreality aliases connected your telephone -- has been gaining steam complete nan past year, and was each complete CES this past January. At GTC, Nvidia made respective investments successful nan technology, ranging from caller models to support for nan information that makes aliases breaks beingness AI systems. 

Here's what's new. 

New models for beingness AI

Nvidia released respective caller instauration models geared towards improving really robots and vehicles usability successful nan existent world. They see Cosmos 3, which generates synthetic worlds to thief beingness AI navigate analyzable environments; Isaac GR00T N1.7, an "open reasoning imagination connection action (VLA) model" built for humanoid robots, which nan institution says is "commercially viable for real-world deployment"; and Alpamayo 1.5, different reasoning VLA exemplary that gives self-driving vehicles amended navigation guidance and punctual specification. 

Also: Nvidia bets connected OpenClaw, but adds a information furniture - really NemoClaw works

Nvidia called Alpamayo 1.5 "a awesome upgrade" wrong its existing autonomous conveyance exemplary family, noting it "takes driving video, ego-motion history, navigation guidance and earthy connection prompts arsenic inputs." It turns those inputs into driving trajectories that fto developers intimately way a vehicle's behaviour and create information guardrails done prompts. Nvidia said Alpamayo 1.5 tin thief return autonomous driving to nan adjacent level by making it easier to study from unpredictable roadworthy events, upwind conditions, aliases pedestrian activity. 

Currently, Nvidia said, its customers are utilizing Cosmos 3 to train beingness AI systems and GR00T N1.7 to "scale humanoid robot deployment." 

Autonomous vehicles 

With nan image of 110 different robots down him, Nvidia CEO Jensen Huang described our present, saying nan "ChatGPT infinitesimal of self-driving cars has arrived." 

Nvidia is broadening its business pinch Uber, saying it will "launch a fleet of autonomous vehicles" powered wholly by Nvidia's Drive AV package successful 28 cities crossed 4 continents by 2028, pinch Los Angeles and San Francisco starting earlier successful 2027. Presumably, that intends users will beryllium capable to book self-driving cars successful nan Uber app connected a overmuch larger scale. 

Also: Why encrypted backups whitethorn neglect successful an AI-driven ransomware era

"This DRIVE Hyperion-powered fleet will pat into NVIDIA Alpamayo unfastened models and nan NVIDIA Halos operating strategy to accelerate nan improvement and deployment of safe, scalable robotaxi services worldwide," nan institution said successful nan release. 

The institution is besides adding respective automakers, including BYD, Hyundai, Nissan, and Geely, to its robotaxi initiative, which already includes GM, Mercedes, and Toyota. Several of those caller summation companies are continuing to usage Nvidia's Drive Hyperion platform, alongside its Alpamayo models, to standard "level 4" conveyance training, aliases nan highest level of automated driving (a afloat functional self-driving car that has fundamentally nary guidance from quality passengers).

Edge AI and abstraction computing

Nvidia is besides moving pinch T-Mobile and Nokia to velocity up beingness AI utilizing AI power entree web (AI-RAN) infrastructure successful distant locations. The institution says this could thief real-world information postulation for beingness AI transverse unconnected, isolated, aliases overcrowded zones utilizing (but without disrupting) 5G connectivity. 

"By turning nan 5G web into a distributed AI machine pinch T-Mobile and Nokia, we're creating a scalable blueprint for nan world's separator AI infrastructure," Huang said successful nan announcement. 

The use of separator AI is debased latency: Local hubs let accusation to move much quickly than erstwhile it has to transverse nan full internet. Nvidia's business uses T-Mobile's existing infrastructure to support that for nan improvement of beingness AI. The institution said inferior and operations companies are already utilizing beingness AI agents, systems, and integer twins crossed this infrastructure for usage cases for illustration optimizing postulation ray timing aliases fixing transmission lines. 

In another announcement, Nvidia besides nodded to abstraction computing. The institution said its caller platforms, including Vera Rubin, are "unlocking a caller era of abstraction innovation, bringing AI compute to orbital information centers (ODCs), geospatial intelligence and autonomous abstraction operations."

Also: What's nan woody pinch beingness AI? Why nan adjacent frontier of tech is already each astir you

What that intends successful practice: Nvidia is connected nan measurement to AI applications that tin run betwixt Earth and space, arsenic good arsenic betwixt abstraction and space. Nvidia said its IGX ThorTM and Jetson OrinTM platforms connection nan energy-efficient conclusion and information processing required to do thing successful orbit -- which is separator AI, functioning arsenic a section hub successful space, extracurricular nan cloud. 

"As we deploy outer constellations and research deeper into space, intelligence must unrecorded wherever information is generated," Huang said successful nan release. 

But orbital information centers are still theoretical -- not impossible, but not yet a afloat reality. While Nvidia's IGX Thor and Jetson Orin platforms are disposable today, nan Vera Rubin Space-1 constituent of nan company's abstraction initiative, announced today, will beryllium "available astatine a later date." 

A caller 'factory' for beingness AI data 

Physical AI lives successful robotics, autonomous vehicles, and different real-world applications, which tin mean higher stakes if thing goes mechanically aliases computationally wrong. That problem is champion avoided pinch high-quality training information that prepares beingness AI systems for arsenic galore situations arsenic imaginable to guarantee they return safer, much predictable, and much effective action. 

To travel its attraction connected beingness AI, Nvidia besides announced its Physical AI Data Factory Blueprint, an "open reference architecture that unifies and automates really training information is generated, augmented and evaluated, reducing nan costs, clip and complexity of training beingness AI systems astatine scale."

Also: Why buying into Moltbook and OpenClaw whitethorn beryllium Big Tech's astir vulnerable stake yet

Set to beryllium disposable adjacent period connected GitHub, Blueprint lets companies usage Nvidia's Cosmos family of world instauration models to process real-world information and generate synthetic data astatine standard to train beingness AI systems. It besides supports reinforcement learning and testing processes for autonomous vehicles and different beingness AI systems. According to Nvidia, Blueprint ensures datasets are divers by including synthetic examples of separator cases and different infrequent scenarios that are harder aliases costly to archive successful nan existent world. 

While it won't beryllium disposable wide until April, Nvidia said Uber is already utilizing Blueprint to create autonomous vehicles, and Skild AI is utilizing it for general-purpose robotics. 

The large picture

Advancements successful beingness AI person user applications, for illustration Waymo cars and nan viral location chore robots you've apt travel across, but are astir instantly applicable to business engineering. More capable, autonomous robots will person nan biggest effect connected our nationalist and business landscapes: connected roads, successful factories, and, evidently, stepping crossed taxable parks.