Cadence unveils Maestro AI models that chipmakers can run in-house, built partly on Nvidia's Nemotron
Cadence Design Systems on Oct. 26 launched Maestro, a portfolio of AI models tuned for chip-design agents that customers can run on any cloud or entirely behind their own firewalls, diverging from rival Synopsys' OpenAI-hosted approach.
Cadence Design Systems, one of the largest makers of chip-design software, on Oct. 26 launched Maestro, a portfolio of domain-specialized AI models built to power its chip and system design agents.
According to Reuters, Maestro was built in part on Nvidia's open-weight Nemotron model and is secured with Nvidia's OpenShell safety technology. Customers can run the models on Cadence's cloud or any public cloud, or keep them entirely in-house behind their own firewalls and data centers, and can fine-tune them on their own chip designs and data.
Cadence says early results show up to 30% higher quality of results and up to five times better token efficiency than a general-purpose model. Tom's Hardware reported that both figures are measured against the unnamed open-source model Maestro started from, and that Cadence did not specify exactly what the 30% figure measures. The first models will reach select early-access customers in the fourth quarter of 2026.
The approach contrasts with that of Cadence's largest rival, Synopsys, which said on Sept. 30 it would develop GPT-Synopsys with OpenAI and share revenue with the ChatGPT maker; that model will run on OpenAI-hosted infrastructure. Both companies are racing to have AI agents operate their tools on behalf of human engineers, speeding up the painstaking work of arranging billions of transistors to meet cost and performance targets.
Paul Cunningham, a Cadence senior vice president, described Maestro as model-agnostic, saying "there are more (AI) model providers out there than there are (chip design software) vendors," and that open- and closed-weight approaches are being "pursued equally." He told Tom's Hardware that agent costs have "really started to explode," making it natural to move more of an agent's iteration "into the weights of the models themselves." Cadence shares rose 3.9% after the announcement.
Sources: Reuters via Investing.com, Tom's Hardware, CNA. Image: Reuters (via CNA).
Related articles
