CEO Jensen Huang confirmed among the robotics builders in NVIDIA’s ecosystem in his GTC 2026 keynote. Supply: NVIDIA
At its annual GPU Know-how Convention, or GTC, NVIDIA Corp. confirmed off its partnerships with the worldwide robotics ecosystem, together with 110 robotic mind builders, industrial automation leaders, and humanoid pioneers. The corporate touted its processing {hardware}, simulation software program, and companion connections powering “production-scale bodily AI.”
NVIDIA this week unveiled new NVIDIA Isaac simulation frameworks and new NVIDIA Cosmos and NVIDIA Isaac GR00T open fashions for the trade to develop, practice, and deploy the following technology of clever robots. Among the many corporations partnering with NVIDIA are ABB Robotics, AGIBOT, Agility, FANUC, Determine, Hexagon Robotics, KUKA, Skild AI, Common Robots, World Labs, and Yaskawa.
“Bodily AI has arrived — each industrial firm will change into a robotics firm,” said Jensen Huang, founder and CEO of NVIDIA. “NVIDIA’s full-stack platform — spanning computing, open fashions and software program frameworks — is the muse for the robotics trade, uniting a worldwide ecosystem to construct the clever machines that can energy the following technology of factories, logistics, transportation and infrastructure.”
Simulation validates the world’s largest robotic fleets, says NVIDIA
As industrial robotics turns into extra AI-driven, producers want bodily correct, high-fidelity simulation to design, take a look at and optimize techniques earlier than deployment, asserted NVIDIA.
“Main industrial software program suppliers are incorporating NVIDIA’s applied sciences,” stated Rev Lebaredian, vp of Omniverse and simulation know-how at NVIDIA. “For robotics, compute is knowledge. The true world is unpredictable, and we will’t seize sufficient human-scale knowledge from demonstration. The bodily AI manufacturing unit blueprint is an open reference … shortening coaching time from years to months.”
With a worldwide set up base exceeding 2 million robots, FANUC, ABB Robotics, Yaskawa, and KUKA are integrating NVIDIA Omniverse libraries and NVIDIA Isaac simulation frameworks into their digital commissioning options. NVIDIA stated they’re utilizing them to develop and validate complicated robotic functions and full manufacturing traces by bodily correct digital twins.
To energy superior intelligence on the manufacturing line, the businesses are integrating NVIDIA Jetson modules into their controllers for real-time AI inference on the edge.
Robotic builders constructing brains for any embodiment
Robotics is evolving from task-specific machines to adaptable generalist-specialist techniques that keep the precision and reliability required for industrial-grade deployment, in keeping with NVIDIA. To attain this, robots want humanlike reasoning and the power to understand, resolve and act autonomously.
Main builders similar to FieldAI and Skild AI are constructing generalized robotic brains utilizing NVIDIA Cosmos world fashions for knowledge technology and Isaac simulation frameworks to validate insurance policies in simulation, enabling any robotic to grasp new duties with minimal retraining.
World Labs is utilizing Isaac Sim to validate its generative world fashions, whereas Generalist AI is utilizing Cosmos to discover producing artificial knowledge.
NVIDIA additionally introduced Cosmos 3.0, which it claimed is the primary world basis mannequin unifying artificial world technology, imaginative and prescient reasoning, and motion simulation to speed up the event of generalized robotic intelligence for complicated environments.
“It unifies imaginative and prescient, reasoning, motion, and simulation,” stated Lebaredian in a press briefing. “The ‘large bang’ of bodily AI has arrived, with $20 billion invested in humanoid robots.”
The most recent Isaac Lab-Enviornment launch simplifies setting composition and accelerates complicated job creation to assist builders consider numerous duties in parallel. Supply: NVIDIA
AI, fashions energy the following technology of humanoids
Constructing humanoid robots is one in every of robotics’ biggest challenges, famous NVIDIA. Replicating human mobility, dexterity, and reasoning requires tightly integrating superior AI, notion and real-time management right into a secure, dependable and autonomous system.
Leaders together with 1X, AGIBOT, Agility, Agile Robots, Boston Dynamics, Determine, Hexagon Robotics, Humanoid, Mentee, and NEURA Robotics are constructing the following technology of humanoids utilizing Cosmos world fashions, Isaac Sim and Isaac Lab to speed up the event and validation of their robots.
NVIDIA additionally launched Isaac Lab 3.0 in early entry, enabling sooner, large-scale robotic studying on NVIDIA DGX-class infrastructure. Constructed on the brand new Newton physics engine 1.0 and the NVIDIA PhysX software program improvement equipment (SDK), it provides multiphysics simulation and improved help for complicated, dexterous manipulation.
AGIBOT, Humanoid, LG Electronics, NEURA Robotics, and Noble Machines are additionally adopting NVIDIA Isaac GR00T N fashions to speed up industrial deployments of their humanoids. NVIDIA stated that GR00T N1.7 is now obtainable in early entry with industrial licensing, bringing generalized robotic abilities together with superior dexterous management to production-ready robotic deployments.
As well as, throughout his GTC keynote, Huang previewed GR00T N2, a next-generation basis mannequin based mostly on DreamZero analysis. Constructed on a brand new world motion mannequin structure, the mannequin may also help robots succeed at new duties in new environments greater than twice as usually as main imaginative and prescient language motion (VLA) fashions.
Slated to be obtainable by the top of the yr, GR00T N2 at present ranks No. 1 on MolmoSpaces and RoboArena for generalist robotic insurance policies, stated NVIDIA.
These techniques are powered by the NVIDIA Jetson Thor robotic computing platform, which the corporate stated permits builders to maneuver from simulation coaching to real-world deployment with higher velocity, intelligence and reliability.
Companions broaden bodily AI to healthcare robotics
Healthcare is a defining alternative for bodily AI however deploying autonomous techniques throughout surgical procedure, imaging and hospitals calls for infrastructure constructed for the very best requirements of security and regulatory rigor.
CMR Surgical is utilizing Cosmos-H simulation to coach and validate robotic intelligence for its Versius surgical system previous to medical deployment.
Johnson & Johnson MedTech is utilizing Isaac Sim- and Cosmos-based post-training workflows to coach and validate techniques for the Monarch Platform for Urology. As well as, Medtronic is exploring NVIDIA IGX Thor to ship mission-critical precision and purposeful security in surgical robotic techniques.
“Jetson Thor is designed to be the engine that drives the brains of humanoids, cell robots, and manipulators, whereas IGX Thor is an embedded PC meant to be stationary, driving bodily AI workloads on the edge,” defined Lebaredian. “IGX Thor gives high-speed sensor processing and purposeful security on the edge, making it appropriate for industries together with building, manufacturing, healthcare, and house, the ultimate frontier.”
NVIDIA acts as a worldwide catalyst for robotic innovation
By constructing an open and built-in platform for designing, coaching, testing and deploying bodily AI, NVIDIA stated it fostering collaboration within the robotics ecosystem, which is crucial to rising adoption. The Santa Clara, Calif.-based firm stated its partnerships with trade leaders are already translating platform integration into real-world affect.
“NVIDIA is the world’s largest robotics firm that doesn’t construct robots,” stated Lebaredian. “We join everybody from trade giants to new startups.”
Skild AI is partnering with industrial automation supplier ABB Robotics and cobot chief Common Robots to deploy its generalized robotic intelligence throughout totally different industries and duties. By embedding a shared intelligence layer into broadly deployed techniques, producers can lengthen automation into extra dynamic and variable functions with out constructing task-specific code for each workflow.
On the similar time, Skild AI is partnering with Foxconn on high-precision meeting for NVIDIA Blackwell manufacturing traces, enabling Foxconn’s AI-driven dual-arm manipulators to grasp the trade’s most complicated manufacturing duties.
NVIDIA cited as one other instance Lightwheel, which is co-developing and calibrating the Newton physics engine to allow Samsung’s meeting robots to grasp intricate cable dealing with in simulation, delivering greater precision and sooner meeting traces.
PTC introduced a brand new robotics design-to-simulation workflow from its cloud-native Onshape computer-aided design (CAD) and product knowledge administration platform to NVIDIA Isaac Sim. It has created a CAD-to-OpenUSD bridge to allow engineering groups like FANUC America Corp. and Fauna Robotics to design and validate their robotic techniques inside bodily correct digital twins.
WORKR is integrating its AI platform with ABB’s industrial robots. It makes use of NVIDIA Omniverse libraries as a part of its WorkrCore to coach a robotic workforce that may be deployed by small and midsize producers in minutes, with out programming data.
KION Group is working with NVIDIA and Accenture to advance autonomous warehouse techniques. It makes use of Omniverse and a bodily AI-powered digital twin and techniques structure pioneered by Accenture. KION engineers can create large-scale, physics-accurate warehouse digital twins to coach and take a look at fleets of Jetson-based autonomous forklifts for GXO, the world’s largest pure-play contract logistics supplier.
Microsoft Azure and Nebius are integrating the NVIDIA Bodily AI Information Manufacturing unit blueprint to allow scalable, agent-driven artificial knowledge technology for his or her builders, together with FieldAI, Teradyne Robotics, Hexagon Robotics on Azure, and RoboForce on Nebius.
CoreWeave is integrating Isaac Lab to construct robotic studying pipelines, whereas Alibaba Cloud is integrating NVIDIA’s complete bodily AI stack into its Platform for AI to speed up end-to-end robotics improvement.
Kamino, a GPU-accelerated physics simulator constructed by Disney on the NVIDIA Warp framework and built-in into Newton, permits the coaching of robotic insurance policies for Disney Imagineering’s Olaf and BDX Droids. It permits Olaf to study to handle its personal warmth and cut back affect noise, and enabling BDX Droids to navigate complicated environments.
In his GTC keynote, Huang was joined by Disney ’s Olaf forward of the robotic’s debut at Disneyland Paris on March 29.
NVIDIA intends to assist launch bodily AI pioneers
NVIDIA stated it’s dedicated to making sure that the instruments for bodily AI are accessible to each innovator, from early-stage startups to the worldwide open-source group.
By way of the NVIDIA Inception program, a worldwide startups incubator with over 40,000 members, NVIDIA gives robotics pioneers with a devoted entry level to its open bodily AI stack. Inception startup members similar to Bedrock Robotics, Dexterity AI, Flexion, Lightwheel, RIVR, Commonplace Bots, Vention, and World Labs get entry to technical steering and high-performance computing sources, in addition to connections to key companions and clients throughout the robotics ecosystem.
NVIDIA has additionally partnered with Hugging Face to combine Isaac and GR00T into the LeRobot open-source framework, connecting NVIDIA’s 2 million robotics builders with Hugging Face’s 13 million AI builders worldwide to speed up the event of open-source robotics.
NVIDIA and Hugging Face are integrating robotics improvement instruments into the LeRobot open-source framework. Supply: NVIDIA
The submit NVIDIA works with international robotics leaders to make bodily AI a actuality appeared first on The Robotic Report.
