Cadence Design Systems has unveiled Cadence Maestro, an AI model product family that supports semiconductor design, Reuters reported on Oct. 26 local time.

Cadence will strengthen cooperation with Nvidia and open-source AI models through Cadence Maestro.

The move differs from rival Synopsys, which recently teamed up with OpenAI.

Semiconductor design involves precisely placing billions of transistors to meet cost and performance targets. Cadence and Synopsys have recently been active in supporting AI agents to operate design tools instead of engineers to speed up work.

Cadence developed Maestro using parts of Nvidia's open-weight model Nemotron.

It also applied Nvidia's OpenSheath technology for security. Customers can run the model in Cadence Cloud or public cloud, and they can also operate it on premises within their own firewall and data center.

Customers can also further train the model with their own semiconductor design data. Cadence said early results showed quality was up to 30 percent higher and cost efficiency up to 5 times higher than general-purpose models.

Earlier, Synopsys said on Sept. 30 it agreed with OpenAI to jointly develop AI models and share revenue. The AI model they are developing, GPT-Synopsys, is based on OpenAI infrastructure.

Paul Cunningham (폴 커닝햄), a senior vice president at Cadence, said there are more AI model companies than semiconductor design software companies. He stressed that Maestro is not tied to any specific model. He said he also has cooperation with commercial AI research and development companies in mind, but gave no details.

The first Maestro models will be supplied first to some customers in the fourth quarter.