AI-modeller skal revolutionere agentbaseret design af chips og systemer
Cadence har introduceret Maestro – en ny portefølje af specialiserede AI-modeller, der er designet til at forbedre ydeevnen af virksomhedens eksisterende AI Super Agents til design af chips og elektroniske systemer (in english).
Cadence has announced the Maestro AI Model Portfolio, which powers Cadence AI Super Agents across semiconductor and system design workflows.
By integrating these models with reinforcement learning (RL) loops and its proven electronic design automation (EDA) and system design automation (SDA) tools, Cadence delivers a deeply integrated agentic engineering platform designed to improve results, accelerate time to market and continuously enhance performance through recursive self-improvement.
The Maestro technology offers customers a secure, Cadence-managed platform that combines exceptional domain-specialised intelligence and expertise with the control, choice, and flexibility they need to innovate with confidence.
The Maestro portfolio spans multiple model sizes and architectures, combining domain-specialised post-trained large language models (LLMs), trained decision models and self-improving agent skills.
Integrated with Cadence AI Super Agents - including AuraStack, ChipStack, InnoStack, and ViraStack AI Super Agents - as well as its proven EDA and SDA tools and reinforcement learning, the Maestro portfolio brings specialised intelligence to complex engineering workflows through an optimal blend of frontier, open-weight and domain-specialised models across cloud and on-premises environments.
- Revolutionising complex engineering tasks requires deep domain-specialised intelligence. Combining our Maestro portfolio of domain-specialised models with leading frontier and open-weight models enables our AI Super Agents to drive Cadence EDA and SDA tools with exceptional skill and efficiency to transform chip and system design, says Anirudh Devgan, president and CEO at Cadence.
- Agentic chip design requires AI that can generate design content, evaluate trade-offs and learn from verified engineering results, addsTim Costa, vice president and general manager of industrial and computational engineering at NVIDIA.
- Our collaboration with Cadence brings NVIDIA Nemotron, OpenShell and deep engineering expertise to Maestro’s domain-specialised models. By grounding reinforcement learning in feedback from EDA tools and deterministic solvers, we can teach agents to make better design decisions based on how well their work meets engineering objectives and constraints.
A self-improving learning environment for engineering AI
Initial results show up to 30% improvement in quality of results and up to 5X greater token efficiency versus a general-purpose foundation model. These gains come from a vertically integrated learning environment combining Cadence domain knowledge, curated engineering datasets, AI Super Agents, EDA and SDA tools, and evaluation harnesses.
Cadence’s principled algorithms and deterministic solvers evaluate outputs against engineering objectives, constraints and ground truth, creating closed and verified feedback loops for reinforcement learning and recursive improvement.
Domain intelligence with intelligent model choice
The Cadence learning environment unifies complementary AI capabilities in a single workflow. Frontier and open-weight models provide broad language, reasoning and multimodal capabilities; the Maestro AI Model Portfolio provides domain expertise for complex chip and system design.
Given a high-level objective, Cadence AI Super Agents interpret design intent, decompose tasks, orchestrate the Maestro portfolio with other leading frontier and open-weight models, and drive Cadence EDA and SDA tools.
The combined solution interprets specifications, generates design content, analyses results, evaluates trade-offs and guides decisions based on engineering requirements, cost and security.
The Maestro portfolio, including NVIDIA Nemotron and other models, can be further post-trained on a customer’s proprietary data, designs, and workflows and, by balancing EDA and SDA tool calls with LLM invocations, it significantly reduces token usage while maintaining execution efficiency.
The complete Cadence agentic environment - including AI Super Agents, Maestro AI Model Portfolio and EDA tools - can run on Cadence’s secure cloud platform, secure public clouds, or in a customer-controlled on-premises environment to meet data privacy, security and deployment requirements.
Cadence and NVIDIA are also working together on agent security through integration with the NVIDIA OpenShell Reference Architecture.
