October 26, 2026 — Cadence Design Systems launched Maestro, a portfolio of AI models built for chip design that customers can run entirely inside their own data centers. Why it matters: it is a direct, privacy-first counter to Synopsys’s rival model being built with OpenAI on OpenAI’s infrastructure.
What happened
Cadence, one of the two companies whose software the semiconductor industry uses to design chips, on Monday launched Cadence Maestro — a suite of AI models tuned for the work of arranging billions of transistors to hit cost and performance targets. The models are meant to power AI agents that use Cadence’s design tools on behalf of human engineers, speeding up a process that now stretches for months on advanced chips.
The details
Maestro was built in part on Nvidia’s open-weight Nemotron model and is secured with Nvidia’s OpenShell safety technology, deepening an already close partnership between Cadence and Nvidia. The deployment terms are the headline: customers can run the models on Cadence’s own cloud, on any public cloud, or entirely in-house behind their own firewalls and data centers — and they can fine-tune the models on their own chip designs, the most closely guarded intellectual property a chipmaker owns. Cadence says early results show up to 30% higher quality and up to five times better cost-efficiency versus general-purpose models. The first Maestro models reach select early-access customers in the fourth quarter.
A split in how chip AI gets sold
The launch draws a clear line against Cadence’s largest competitor. 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. Cadence is betting the opposite way. Senior vice president Paul Cunningham framed Maestro as model-agnostic, noting there are more AI model providers than chip-design software vendors, and said working with commercial labs is part of the Maestro vision — but the architecture lets customers keep their designs, and their model, at home.
Why it matters
Chip designs are among the most sensitive files in technology, and every AI vendor selling into semiconductors has to answer the same question: does our crown-jewel IP have to leave the building? Cadence’s answer is no, and it is the same argument driving on-premise AI across finance, defense, and healthcare. If Maestro’s early quality and cost claims hold up with real customers, expect “runs behind your firewall” to become a standard demand in enterprise AI contracts — not a premium feature.
FAQ
What is Cadence Maestro?
A portfolio of AI models from Cadence, tuned specifically for designing computer chips, launched October 26, 2026.
What is Maestro built on?
Partly on Nvidia’s open-weight Nemotron model, secured with Nvidia’s OpenShell technology.
When can customers use it?
The first models reach select early-access customers in the fourth quarter of 2026.
Sources: Reuters