Can AI agents verify their own engineering decisions? A validation-driven workflow is aiming to improve trust across chip and PCB design.
Siemens has expanded its collaboration with NVIDIA to introduce self-verifying agentic artificial intelligence (AI) workflows for semiconductor and printed circuit board (PCB) design. The capability combines AI reasoning with deterministic, physics-based electronic design automation (EDA) tools, allowing autonomous agents to continuously validate their outputs rather than relying solely on generated results.
The development addresses a key challenge in applying AI to engineering workflows: ensuring that AI-generated decisions are accurate and reproducible. By validating outputs against established EDA engines, the approach aims to improve design quality, shorten engineering cycles and increase confidence in long-running AI-assisted design tasks. The self-verifying workflow also integrates with Siemens’ Intelligence Center X, enabling AI agents to coordinate across design, manufacturing and supply chain operations.
The enhanced Fuse EDA AI Agent incorporates NVIDIA technologies, including NeMo Gym, OpenShell, Nemotron reasoning models, Switchyard and accelerated computing based on CUDA-X libraries. According to the companies, these additions improve tool-calling reliability, token efficiency and result quality while enabling secure multi-agent collaboration across semiconductor and PCB design flows. The platform supports workflows spanning high-level synthesis, verification, custom IC design, physical implementation, signoff verification, design-for-test, 3D IC integration and PCB design.
The collaboration also extends agentic AI capabilities to Siemens’ Solido Characterization Suite, automating Liberty file generation and verification for advanced-node standard cells, memory and custom IP libraries. The company says the workflow reduces characterization turnaround time by more than ten times while lowering token costs by five to ten times. Additional capabilities have also been introduced for layout analysis and digital verification to help identify design issues earlier in the development process.
“Our expanded collaboration with NVIDIA enhances our domain-specific industrial AI, physics-based EDA engines and accelerated computing to create trusted, self-verifying AI workflows,” says Amit Gupta, Senior Vice President and Chief AI Strategy Officer, Siemens EDA, Siemens Digital Industries Software. “By enabling autonomous and long-running EDA AI agents to continuously validate their decisions against proven engineering tools, we help customers accelerate development, improve design quality and increase confidence in trusted engineering outcomes.”
