Major semiconductor design software developer Synopsys has officially announced the AgentEngineer solution suite, featuring specialized agents for long-horizon tasks built on its new Autopilot artificial intelligence platform. The package spans six key chip development domains: chip verification, full-system validation, physical implementation, analog and mixed-signal, as well as modeling and analysis.
The system features a three-tier architecture comprising specialized super-agents that coordinate worker agents, worker agents that address distinct engineering tasks directly, and tool-level engines that execute requested work without making decisions or setting goals. Agents designed for long-horizon tasks help solve the complexity problem by executing workflows requiring hundreds or thousands of reasoning steps.
Security and Architecture of Autonomous Systems
The platform covers all development stages from AI agent coordination to telemetry collection, anchored by a cognitive model providing context intelligence. Intellectual property security is ensured through access control systems, encryption, and secure execution environments. Clients can deploy commercial, open-source, or fine-tuned AI models within the Synopsys cloud, a private cloud, or local infrastructure.
Real-World Results and Performance Metrics
During preliminary rollout, joint chip verification testing with Nvidia accelerated verification completion by up to 50 times while expanding verification completeness by 20%. Fujitsu reported up to a 30% increase in productivity for RTL code generation tasks. General availability is scheduled for late 2026, with components meeting the highest L5 autonomy level criteria.
