Synopsys has introduced AgentEngineer, a portfolio of domain-specific AI agents built on its new Autopilot Platform. The agents target verification, system validation, implementation, analog and mixed-signal design, manufacturing, and simulation and analysis. According to Synopsys, more than 50 customer engagements are underway, and general availability is planned for the end of 2026.
The platform is organized in three layers: long-horizon AgentEngineers that orchestrate task agents, task agents that handle bounded engineering tasks, and a tool layer that executes work without setting goals. Synopsys says the agents can plan, check intermediate results, and course-correct, but human approval checkpoints remain. Customers can bring their own language models, data, and infrastructure, and deploy on Synopsys Cloud, their own cloud, or on-premises.
Synopsys cites performance gains such as up to 50x faster verification closure and 20% higher coverage, but those figures were previously reported from July work with Nvidia. A 30% productivity boost is at the top of a range Fujitsu reported, and a 2x token-efficiency claim comes from an unnamed customer. AheadComputing reported reduced manual effort without giving a number, while Intel, MediaTek, and Samsung endorsed the technology without hard results.