Cadence unveils Nvidia-backed AI models for in-house chip design

Key facts

• Cadence Maestro AI models can be run in-house or on cloud infrastructure.
• Models can be fine-tuned on customer chip designs.
• Early results show up to 30% higher quality and five times better cost-efficiency.
• Cadence's approach contrasts with Synopsys's deal with OpenAI.

Cadence Design Systems on Monday launched Maestro, a new suite of AI models designed to assist in the complex process of computer chip design. The models, developed in partnership with Nvidia, leverage Nvidia's Nemotron model and OpenShell safety technology. Cadence highlighted that these AI tools can be deployed on-premises, within private data centers, or on cloud infrastructure, allowing chipmakers to retain control over their data and designs. Early users have reported significant improvements in design quality and cost-efficiency compared to general-purpose AI models. This offering positions Cadence against its main competitor, Synopsys, which has pursued a different strategy by partnering with OpenAI for its AI chip design model.