At CadenceLIVE Silicon Valley 2026, Cadence announced an expanded collaboration with NVIDIA aimed at transforming engineering design through agentic AI, physics-based simulation, and digital twins. The partnership targets faster innovation across semiconductor design, physical AI systems, and hyperscale AI factories.
By integrating Cadence’s electronic design automation (EDA) and system design technologies with NVIDIA’s CUDA-X platform, AI physics, and Omniverse libraries, the companies aim to significantly accelerate complex engineering workflows. The combined solution is designed to deliver up to 100x speed improvements in simulation and analysis, enabling faster and more accurate product development.
A key highlight is Cadence’s new AgentStack, an agentic AI system that orchestrates chip and system design processes. Building on its ChipStack AI Super Agent, AgentStack extends AI-driven automation across the entire design flow—from RTL to system-level development—reducing iteration cycles from days to hours. NVIDIA is already adopting the platform internally, signaling strong industry validation.
The collaboration also advances physical AI development by integrating Cadence’s simulation tools with NVIDIA Isaac robotics platforms. This enables end-to-end workflows—from AI training and physics modeling to real-world deployment—helping close the “simulation-to-reality” gap in robotics and autonomous systems.
In AI infrastructure, the partnership introduces digital twin solutions for AI factories using NVIDIA Omniverse. These virtual models allow operators to optimize performance metrics such as “tokens per watt,” a key efficiency benchmark for large-scale AI systems. Early simulations show potential gains of up to 32% in efficiency, translating into significant cost savings and increased revenue at scale.
The expanded alliance underscores a broader industry shift toward AI-driven, simulation-first engineering—where digital twins and agentic AI redefine how complex systems are designed, tested, and deployed.
