Nvidia partners with Cadence

Nvidia and Cadence said on April 15 they are expanding their partnership to build more realistic virtual testing grounds for robots before those machines move into factories and warehouses. (reuters.com) Robot simulation is software that lets companies train and test machines in a digital copy of the real world, instead of learning every task on a physical robot. Nvidia’s Isaac Sim is built for that job, using physics models and synthetic sensor data to mimic cameras, lidar, and motion in virtual environments. (developer.nvidia.com, github.com) Cadence makes engineering software that predicts how physical systems behave, including motion, heat, and other forces that matter when a robot touches or moves through the world. The new tie-up combines those physics and digital-twin tools with Nvidia’s robotics software stack, the companies said at CadenceLIVE Silicon Valley 2026 in San Jose. (cadence.com, reuters.com) The immediate problem is the “simulation-to-real” gap: a robot can perform well in software and still fail when lighting changes, surfaces slip, or objects bend in ways the model did not capture. Cadence and Nvidia said higher-fidelity simulation should improve training, testing, and validation before deployment. (thenextweb.com, nvidia.com) Nvidia has been pushing deeper into what it calls physical artificial intelligence, or software that learns to act in the physical world rather than just generate text or images. Reuters reported that Chief Executive Jensen Huang and Cadence Chief Executive Anirudh Devgan announced the robotics effort together on Wednesday, April 15. (reuters.com) The companies described a broader expansion than robotics alone. Cadence said the partnership also covers agentic artificial intelligence, semiconductor design, and digital twins for large-scale computing infrastructure, with robotics as one of the clearest near-term use cases. (cadence.com, forbes.com) That fits Nvidia’s recent robotics push. The chip company has also announced industrial robotics work with Fanuc, pairing robot hardware with Nvidia computing and simulation tools for factory automation. (assemblymag.com, reuters.com) For Cadence, the deal extends software long used by chip and system designers into robot training, where the sales pitch is fewer costly real-world tests and faster iteration. For Nvidia, it adds more detailed physics to the virtual worlds its customers already use to train and validate machines. (reuters.com, cadence.com) The bet is simple: if the digital world behaves more like the real one, robots should need less retraining after deployment. That would make simulation not just a design tool, but a larger part of how robots get built, tested, and shipped. (thenextweb.com, nvidia.com)