Nvidia wants to be the Android of generalist robotics 

Nvidia released a caller stack of robot instauration models, simulation tools, and borderline hardware astatine CES 2026, moves that awesome the company’s ambition to go the default level for generalist robotics, overmuch arsenic Android became the operating strategy for smartphones. 

Nvidia’s determination into robotics reflects a broader manufacture displacement arsenic AI moves disconnected the unreality and into machines that tin larn however to deliberation successful the carnal world, enabled by cheaper sensors, precocious simulation, and AI models that progressively tin generalize crossed tasks. 

Nvidia revealed details connected Monday astir its full-stack ecosystem for carnal AI, including caller unfastened instauration models that let robots to reason, plan, and accommodate crossed galore tasks and divers environments, moving beyond constrictive task-specific bots, each of which are disposable connected Hugging Face. 

Those models include: Cosmos Transfer 2.5 and Cosmos Predict 2.5, 2 satellite models for synthetic information procreation and robot argumentation valuation successful simulation; Cosmos Reason 2, a reasoning imaginativeness connection exemplary (VLM) that allows AI systems to see, understand, and enactment successful the carnal world; and Isaac GR00T N1.6, its next-gen imaginativeness connection enactment (VLA) exemplary purpose-built for humanoid robots. GR00T relies connected Cosmos Reason arsenic its brain, and it unlocks whole-body power for humanoids truthful they tin determination and grip objects simultaneously. 

Nvidia besides introduced Isaac Lab-Arena astatine CES, an open-source simulation model hosted connected GitHub that serves arsenic different constituent of the company’s carnal AI platform, enabling harmless virtual investigating of robotic capabilities.

The level promises to code a captious manufacture challenge: arsenic robots larn progressively analyzable tasks, from precise entity handling to cablegram installation, validating these abilities successful carnal environments tin beryllium costly, slow, and risky. Isaac Lab-Arena tackles this by consolidating resources, task scenarios, grooming tools, and established benchmarks similar Libero, RoboCasa, and RoboTwin, creating a unified modular wherever the manufacture antecedently lacked one.

Supporting the ecosystem is Nvidia OSMO, an open-source bid halfway that serves arsenic connective infrastructure that integrates the full workflow from information procreation done grooming crossed some desktop and unreality environments. 

And to assistance powerfulness it all, there’s the caller Blackwell-powered Jetson T4000 graphics card, the newest subordinate of the Thor family. Nvidia is pitching it arsenic a cost-effective on-device compute upgrade that delivers 1200 teraflops of AI compute and 64 gigabytes of representation portion moving efficiently astatine 40 to 70 watts. 

Nvidia is besides deepening its partnership with Hugging Face to fto much radical experimentation with robot grooming without needing costly hardware oregon specialized knowledge. The collaboration integrates Nvidia’s Isaac and GR00T technologies into Hugging Face’s LeRobot framework, connecting Nvidia’s 2 cardinal robotics developers with Hugging Face’s 13 cardinal AI builders. The developer platform’s open-source Reachy 2 humanoid present works straight with Nvidia’s Jetson Thor chip, letting developers experimentation with antithetic AI models without being locked into proprietary systems.  

The bigger representation present is that Nvidia is trying to marque robotics improvement much accessible, and it wants to beryllium the underlying hardware and bundle vendor powering it, overmuch similar Android is the default for smartphone makers.

There are aboriginal signs that Nvidia’s strategy is working. Robotics is the fastest increasing class connected Hugging Face, with Nvidia’s models starring downloads. Meanwhile robotics companies from Boston Dynamics and Caterpillar to Franka Robots and NEURA Robotics are already utilizing Nvidia’s tech.