Nvidia expands into robotics simulation
Nvidia widened a robotics partnership with Cadence to integrate AI models and simulation engines for robot training. (ts2.tech) The report frames the work as tying accelerator vendors into EDA and simulation workflows used for embodied‑system development. (ts2.tech)
Nvidia and Cadence have widened their partnership to connect robot training software with higher-fidelity engineering simulation tools. (cadence.com) Cadence said on April 15 that the expanded deal links Nvidia Isaac Sim and Isaac Lab with Cadence physics models and its VTD and VTDx scenario simulators. Cadence announced the work at CadenceLIVE Silicon Valley 2026 in San Jose. (cadence.com) In plain terms, robot developers first train software in a virtual world, then test it against more detailed models of motion, sensors, and environments before putting machines in the field. Nvidia describes Isaac Sim as a platform for simulation, testing, and synthetic data generation in physically based virtual environments. (developer.nvidia.com) Cadence’s role is the engineering side of that stack: the company sells design software and multiphysics tools that model heat, motion, electromagnetics, and other real-world effects. The new workflow pushes those models into robotics, where bad simulation can leave robots unprepared for factory floors, warehouses, or roads. (cadence.com) Cadence said the combined workflow spans virtual training in Isaac Sim and Isaac Lab, evaluation through Cadence physics models, and mission-scale scenario simulation in VTD and VTDx. Reuters reported the companies described the effort as a way to advance artificial intelligence for robots. (cadence.com) (msn.com) The companies also tied the robotics work to a broader push across semiconductors, physical artificial intelligence systems, and data-center infrastructure. Cadence said its software will use Nvidia CUDA-X libraries, Omniverse, and AI physics models across those engineering workflows. (cadence.com) Cadence said the collaboration can deliver engineering workflow speedups of up to 100 times, though that figure came from the companies and was not broken out specifically for robotics in the release. The same announcement named customers and partners including Honda Research and Development, Samsung, SK Hynix, Ascendence, and Argonne National Laboratory. (cadence.com) The deal also shows how Nvidia is moving beyond selling graphics processors into the software used to design and validate machines that run on them. Cadence’s software has long been central in electronic design automation, the toolchain chip companies use to design and verify semiconductors. (automation.com) (forbes.com) That matters in robotics because training data is expensive, physical testing is slow, and failures in the real world can damage equipment or halt production lines. Cadence and Nvidia are betting that more detailed digital twins, backed by faster accelerated computing, can move more of that work into simulation before a robot touches the real world. (thenextweb.com) (docs.isaacsim.omniverse.nvidia.com) The immediate next step is adoption: Cadence has laid out the software path, and Nvidia already has Isaac tools in market. The test now is whether robot makers use that combined stack as a standard route from virtual training to deployment. (developer.nvidia.com) (cadence.com)