Cadence Design Systems Inc. announced an expansion of its collaboration with NVIDIA Corp. to develop accelerated engineering solutions for chip and system design using artificial intelligence.

The partnership introduces what the companies call agentic AI design solutions that incorporate autonomous agents for translating design intent into automated workflows. These solutions integrate Cadence's chip and system design portfolio with NVIDIA's accelerated computing stack.

The collaboration includes the Cadence Millennium M2000 Supercomputer, which uses NVIDIA Grace CPUs and Blackwell GPUs. According to the companies, this system delivers up to 80 times greater throughput and up to 20 times lower power consumption compared to previous solutions.

Cadence plans to make several accelerated solutions available in 2026, including electronic design automation tools, system design automation platforms, and life sciences applications. The Cadence Clarity 3D Solver configured with NVIDIA RTX Pro 6000 GPU servers demonstrated up to 5 times faster performance compared to equivalent CPU-based solutions.

"The fusion of agentic AI and physics-based design is transforming how the world's most advanced chips are engineered," said Anirudh Devgan, president and CEO of Cadence.

Jensen Huang, founder and CEO of NVIDIA, stated that the companies have created the Millennium M2000 to address the complexity of designing next-generation infrastructure.

Several companies are already implementing these solutions. Honda is using Cadence Fidelity CFD Software on the Millennium M2000 system for turbofan engine simulation. Micron is integrating GPU-accelerated Cadence technologies into its HBM memory design flow. Larsen & Toubro Semiconductor is using the Cadence Spectre X Simulator with NVIDIA GPUs for AI and data center chip design.