Cadence And NVIDIA Expand Partnership To Close Robotics' Sim-To-Real Gap

Cadence Design Systems and NVIDIA have widened their alliance to fuse Cadence's high-fidelity multiphysics engines with NVIDIA Isaac, Cosmos, and Jetson, targeting agentic AI workflows for chip design, robots, and AI factories.

Cadence And NVIDIA Expand Partnership To Close Robotics' Sim-To-Real Gap

Two engineers in front of a screen showing a physics simulation

At CadenceLIVE Silicon Valley 2026, Cadence Design Systems and NVIDIA announced an expanded partnership to deliver accelerated tools across agentic AI, physics-based simulation, and digital twins for semiconductor design, robotics, and AI factories. The deal extends a multi-year collaboration that already spans EDA, multiphysics, and accelerated computing.

Physical AI Stack Meets Isaac And Cosmos

The headline addition is a new joint workflow for physical AI. Cadence's Physical AI Stack — including its high-fidelity multiphysics solvers and Virtual Test Drive (VTD/VTDx) scenario simulation — will integrate with NVIDIA Isaac Sim, Isaac Lab, and the Cosmos open-world models. The companies say the combination gives robot developers an end-to-end loop: world model training on Cosmos, mission-scale scenarios in VTD, precise physics in Cadence, and on-robot deployment on NVIDIA Jetson.

Why It Matters

Robotics teams routinely hit a sim-to-real cliff when policies trained in fast game engines fail on real hardware because the simulator simplified contact, materials, or sensing. Cadence brings decades of multiphysics fidelity that has been used to design satellites and semiconductors; NVIDIA brings the world models, GPU acceleration, and edge inferencing. Combined, the stack aims to compress the iteration loop that has slowed humanoid and autonomous-machine deployments.

Agentic AI For Engineering

Beyond robotics, the partnership embeds NVIDIA's agentic AI frameworks into Cadence design tools, so engineers can task agents with verification, optimization, and trade studies that previously required days of manual setup. Cadence said it will also use NVIDIA Blackwell-class compute to run far larger semiconductor and aerospace simulations than current workflows allow.

Customer Pull

Early adopters include hyperscalers retooling AI factories and robotics OEMs preparing humanoid programs. The companies framed the announcement as part of a broader push to make digital twins a default deliverable for every system, from chips to robots to power plants.

What's Next

Expect tighter packaging with the Cadence Optimality stack and additional joint releases at NVIDIA's COMPUTEX/GTC Taipei keynote, including new Cosmos-aware verification tools. The collaboration follows a string of Cadence-NVIDIA tie-ups dating back to 2023.

Related: Kawasaki's Silicon Valley Physical AI Center, NVIDIA-IREN 5GW AI factory deal.

Reporting based on coverage from Cadence Newsroom, Business Wire, DIGITIMES, and Futurum Group.

Category: Partnerships

Tags: Physical AI world simulation robotic simulation Partnership ai robotics Digital Twin

Related Articles