Nvidia expanded its partnership with chip software maker Cadence to advance AI capabilities in robotics and accelerate the path from simulation to real-world deployment.
Unveiled by the two companies’ CEOs at a Cadence conference in Santa Clara, California this week, the expanded deal will focus on agentic AI, physics-based simulation and digital twins.
At its core, the partnership targets one of robotics’ most persistent challenges: effectively translating robotic learning from simulations to real-world scenarios (the sim-to-real gap).
While Cadence is best known as a supplier of software used in advanced computing chips, the vendor also develops physics simulation engines that predict how materials interact and behave in real-world applications.
Typically used in sectors such as aerospace and automotive design, under the deal, these high-fidelity simulation engines will be integrated with Nvidia’s AI robot training models (including its Isaac open source simulation libraries and Cosmos open world models), with the goal of more precise robot training data.
Specifically, the integration, the companies hope to develop sharper navigation and object-handling capabilities for AI-powered robots.
The collaboration will also extend into AI infrastructure. Cadence will use Nvidia’s Omniverse DSX blueprint to build digital twins of large-scale AI data centers, enabling customers to simulate and optimize facility design before committing to physical deployment.
“Agentic AI and digital twins are reshaping the entire engineering landscape — from semiconductor design to planetary‑scale AI systems,” Anirudh Devgan, president and CEO at Cadence said in a release. “Our expanded collaboration with Nvidia accelerates the convergence of design and physical realization … deliver[ing] unprecedented speed, accuracy and trust in simulation and system development.”
Alongside the infrastructure push, Cadence launched a new agentic AI platform, AgentStack, to take on some of the tasks typically reserved for human engineers, such as system design workflows.
Nvidia will serve as an early adopter, deploying AgentStack across its own design flows and feeding results back to refine the platform.
“Together, Nvidia and Cadence are transforming how engineers design, build and operate the world,” Jensen Huang, Nvidia’s CEO, said in the release. “For the first time, we can innovate in the digital world — exploring, testing and optimizing ideas at unprecedented speed and scale by building everything as full-fidelity digital twins first.”
The Nvidia deal comes just a day after Cadence entered into a new collaboration with Google, integrating Google’s Gemini models with Cadence’s ChipStack AI Super Agent to bolster automation capabilities.