🧠 Food for thought

Implications, context, and why it matters.

This partnership builds on Cadence’s AI-driven momentum

• The collaboration follows Cadence’s run of AI-led gains, with its AI product portfolio contributing to strong 2025 financial results.
• Its intellectual property (IP) business grew more than 25% year over year in the second quarter of 2025, helped by AI plus high-performance computing (HPC) demand and technologies such as High Bandwidth Memory 4 (HBM4) and 224G SerDes, a high-speed chip-to-chip communications technology 1.
• System Design and Analysis revenue rose 35% year over year in the same quarter 1.
• This backdrop helps explain the closer ties, with Nvidia adopting Cadence’s new AgentStack in semiconductor and system design workflows, as stated in the partnership announcement 2.

The collaboration extends Cadence into broader AI systems and AI factories

• The deal tracks Cadence’s move beyond electronic design automation (EDA) software into system-level work that spans semiconductors, physical AI systems, and hyperscale AI factories.
• Cadence plans to tie its simulation plus AI workflows to Nvidia Isaac, an open-source robotics simulation library, and Cosmos, Nvidia’s open-world AI models. Cadence’s CEO said physical AI could outgrow AI infrastructure, with sciences AI after that 1.
• The companies also plan AI factory digital twins, virtual models of real-world facilities, using metrics such as “tokens per watt.” One example models a 10-megawatt AI factory where reduced-power (MaxQ) operation delivered up to 17% more tokens per watt 2.
• They describe this work as a way to test configuration choices plus cooling tradeoffs before deployment, not as a pitch tied to the market for individual software licenses 2.