Siemens AG (ETR: SIE) and NVIDIA Corporation (Nasdaq: NVDA) have expanded their strategic partnership to introduce self-verifying artificial intelligence agents into semiconductor and printed circuit board (PCB) design, marking the latest push to apply autonomous AI to one of the electronics industry’s most computationally intensive engineering workflows.
The companies said the enhanced collaboration integrates NVIDIA’s AI infrastructure, reasoning models and accelerated computing technologies into Siemens’ recently launched Fuse EDA AI Agent platform, enabling AI agents to autonomously execute long-running semiconductor design tasks while continuously validating decisions against deterministic, physics-based electronic design automation (EDA) tools.
The move reflects the semiconductor industry’s growing focus on “agentic AI,” in which AI systems move beyond assisting engineers with isolated tasks to independently orchestrating complex workflows across chip design, verification and implementation while ensuring engineering accuracy through continuous validation.
Unlike conventional generative AI tools, which can produce inconsistent outputs, Siemens said its latest approach combines AI reasoning with trusted EDA engines, allowing autonomous agents to verify every design decision against established engineering models before proceeding.
The updated Fuse EDA AI Agent is also being integrated into Siemens’ Intelligence Center X, the company’s industrial AI platform designed to coordinate engineering, manufacturing and supply chain operations. Siemens said the integration extends its Digital Twin strategy by enabling AI-driven workflows across the semiconductor development lifecycle while maintaining traceability and engineering confidence.
“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,” said Amit Gupta, Senior Vice President and Chief AI Strategy Officer at Siemens EDA, Siemens Digital Industries Software.
He said enabling autonomous AI agents to continuously validate their work against proven engineering tools would help semiconductor companies accelerate product development while improving design quality and confidence in engineering outcomes.
NVIDIA said the increasing complexity of advanced semiconductor devices requires AI systems capable of understanding engineering trade-offs rather than simply generating responses.
“Semiconductor and PCB design are among the most complex engineering challenges in the world, and AI agents need trusted tools to reason, act and verify their work,” said Timothy Costa, Vice President and General Manager of Computational Engineering at NVIDIA.
Costa said the collaboration combines Siemens’ EDA software with NVIDIA’s accelerated computing platform, AI software and Nemotron open models to improve design quality while reducing engineering turnaround times.
The expanded platform incorporates several NVIDIA technologies, including the NeMo Gym open library for developing engineering AI agents, the OpenShell secure runtime environment, Nemotron reasoning models, CUDA-X libraries and GPU-accelerated computing.
According to Siemens, AI agents developed using NeMo Gym can progressively improve over time by learning from previous engineering projects, allowing them to optimise workflow execution, retain engineering context and continuously refine design strategies across successive semiconductor programmes.
The OpenShell runtime is intended to allow enterprise engineering organisations to deploy autonomous AI agents securely across large EDA environments while maintaining governance, auditability and access controls required for commercial semiconductor development.
The companies also said NVIDIA’s latest Nemotron reasoning models enable AI agents to analyse complex engineering trade-offs more efficiently while improving token utilisation during long-running design sessions.
Meanwhile, NVIDIA’s accelerated computing platform and CUDA-X libraries will power both AI reasoning and Siemens’ deterministic EDA engines, reducing simulation and verification workloads from days to hours while maintaining signoff-level engineering accuracy.
The companies said the combined technologies support secure multi-agent collaboration, real-time engineering feedback and faster convergence across semiconductor and PCB design workflows.
Rather than targeting a single stage of development, Siemens’ updated AI platform spans the full semiconductor design flow. The Fuse EDA AI Agent can orchestrate workflows across high-level synthesis using Catapult, digital verification through Questa One and Veloce, custom IC development with Solido, physical implementation using Aprisa, signoff verification through Calibre, design-for-test with Tessent, advanced 3D IC integration using Innovator3D IC and PCB development through Xpedition.
By coordinating specialised AI agents across multiple engineering domains, Siemens aims to automate entire design flows instead of individual engineering tasks, an increasingly important capability as chipmakers adopt heterogeneous architectures, chiplets and advanced packaging technologies.
The collaboration also introduces new AI capabilities into Siemens’ Solido Characterization Suite, where autonomous agents automate one of the most resource-intensive stages of semiconductor development by generating and verifying Liberty timing libraries.
The workflow combines Solido Characterizer with Solido LibSPICE, Solido Generator and Solido Analytics to automate characterisation for advanced-node standard cells, memory devices and custom intellectual property libraries.
Siemens said the system reduces library characterisation turnaround times by more than tenfold while lowering AI token costs by between five and ten times compared with previous workflows.
The companies also introduced Solido Layout Analyzer, an AI-powered tool that allows engineers to analyse layout-dependent effects and parasitic behaviour in post-layout semiconductor designs using natural language prompts. Integrated with Fuse EDA AI Agent, the software can generate reports, recommend corrective actions and identify design issues earlier in the development cycle.
STMicroelectronics N.V. (NYSE: STM) said it plans to evaluate the technology within its non-volatile memory design operations.
“For STMicroelectronics’ non-volatile memory team, the ability to relate layout insight directly to electrical behavior is critical,” said Gianbattista Lo Giudice, Non-Volatile Memory Design Manager at STMicroelectronics.
He said bringing layout-dependent analysis earlier into the design process could improve engineering confidence while reducing debugging time for complex design blocks by several weeks.
The partnership also targets one of the semiconductor industry’s largest productivity challenges—digital verification, which Siemens estimates accounts for as much as 70% of overall chip development time.
Building on its previously announced Questa One Agentic Toolkit, Siemens is integrating NVIDIA’s Nemotron 3 Ultra reasoning model to enable AI agents to evaluate engineering trade-offs while continuously validating designs against golden reference test harnesses.
The company said the technology will help engineering teams detect design flaws earlier, shorten verification cycles and accelerate production readiness.
“We’re at an inflection point where the complexity of AI chips, chiplets and 3D ICs has outpaced traditional verification methodologies,” said Abhi Kolpekwar, Senior Vice President and General Manager of Digital Verification Technologies at Siemens EDA.
He said agentic AI provides a scalable approach capable of orchestrating verification across multiple engineering domains while evaluating billions of verification scenarios within timeframes that were previously unattainable.
The expanded AI-driven EDA capabilities will be introduced in forthcoming releases of Siemens’ AI-native EDA portfolio.
