Cadence Design Systems on Monday launched Maestro, a portfolio of artificial intelligence models tuned to help design computer chips, deepening its partnership with Nvidia and setting up a direct contrast with rival Synopsys's recent deal with OpenAI. The launch, reported by Reuters from San Francisco, lands as the two companies that dominate chip-design software race to put AI agents to work inside the tools engineers use to arrange billions of transistors into working processors.
The new suite is built in part on Nvidia's open-weight Nemotron model and secured with Nvidia's OpenShell safety technology, the companies said. It is aimed at a very specific kind of AI work: not chatting with consumers, but operating the complex, precision software used to lay out a chip, check it for errors, and squeeze more performance out of every square millimetre of silicon.
Designed to run behind the firewall
The most important detail of Maestro may be where it runs. Cadence said customers can use the models on its own cloud, on any public cloud, or entirely in-house, behind their own firewalls and inside their own data centres. They can also fine-tune the models on their own chip designs, rather than sending their work to a model hosted by someone else.
That matters because a chip design is among a company's most valuable secrets. Handing unfinished designs to an outside-hosted AI service is a step many chipmakers have been reluctant to take, even as they face pressure to design faster. Keeping the model in-house removes one of the biggest objections to letting AI near the design process at all.
Cadence said early results show up to 30 percent higher quality and up to five times better cost-efficiency than general-purpose models.
A different bet from Synopsys and OpenAI
Maestro also draws a clear dividing line in the chip-design software market. On September 30, Synopsys said it would develop an AI model with OpenAI and share revenue with the ChatGPT maker. That model, GPT-Synopsys, will run on OpenAI-hosted infrastructure, meaning customers' design work would flow through OpenAI's systems.
Cadence is betting the opposite way. Paul Cunningham, a Cadence senior vice president, framed Maestro as model-agnostic, noting that there are more AI model providers than there are chip-design software vendors. Asked whether Cadence might strike its own deals with proprietary AI labs, he said both open-weight and closed-weight approaches are being pursued equally, and that working with commercial labs is part of the Maestro vision, though he declined to elaborate.
The distinction could shape how quickly chipmakers adopt AI design tools. A hosted model offers convenience and the backing of a frontier AI lab. An in-house, model-agnostic approach offers control, and the freedom to swap in a better model later without rebuilding the whole workflow.
What comes next
The first Maestro models will reach select early-access customers in the fourth quarter, Cadence said. Their reception will be closely watched, not least because Cadence reports its own third-quarter results after the US market closes on Monday, with investors looking for signs that AI tools are already contributing to the business.
Whoever wins the race to automate chip design will influence the pace of everything built on chips, from phones and cars to the data centres powering the AI boom itself. Cadence's answer is that the winning model will be the one chipmakers are actually willing to let inside their most guarded rooms.



