Siemens Deploys Agentic AI for Chip Design

Siemens' introduction of agentic AI into its Questa One platform is part of a broader company strategy called Fuse EDA AI. This system-wide initiative integrates both generative and agentic AI across Siemens' entire Electronic Design Automation (EDA) portfolio, aiming to enhance everything from IC design to printed circuit board (PCB) development. The platform is designed to be secure and deployment-ready, allowing customers to use their own data within their own data centers, either on-premises or in the cloud. The AI agents in Questa One are designed to automate and accelerate the verification process, a major bottleneck in chip development. This is a response to a growing industry crisis: the rate of first-silicon success has plummeted from 32% in 2020 to just 14% in 2024, while the percentage of projects running behind schedule has increased from 67% to 75% in the same period. By automating tasks like generating testbenches and creating structured test plans, the AI aims to make the verification process more intelligent and proactive. The specific target of this new AI integration is the Register-Transfer Level (RTL) sign-off. RTL is a level of abstraction used in hardware description languages like Verilog and VHDL to model the flow of data between registers and the logical operations performed on that data. The RTL sign-off phase is critical for ensuring the design is structurally sound and ready for the subsequent steps of synthesis and physical implementation. This move reflects a larger trend of applying AI to complex EDA challenges, with projections suggesting that by 2027, up to 90% of advanced chips will be designed with the help of agentic AI. Competitors like Cadence are also heavily invested in evolving from optimization AI to more autonomous agentic AI systems. The goal is to create a new paradigm sometimes called Agentic Design Automation (ADA), where autonomous agents paired with specialized tools automate labor-intensive engineering loops. Under the hood, Siemens is leveraging partnerships, including with NVIDIA, to use NIM microservices and Nemotron reasoning models to power these new capabilities. The system utilizes a family of specialized parsers to ingest and vectorize various EDA data formats, creating a centralized data lake that helps break down silos between design teams. This allows the AI to perform tasks like smart regression, which uses historical data to intelligently select and prioritize tests. Siemens had previously integrated AI into its EDA tools, such as the Solido Design Environment, which uses machine learning to speed up simulations for mixed-signal ICs. The Questa One platform itself is an evolution of the Questa Verification IQ tool, expanding on its use of analytical, predictive, and now generative and agentic AI. The system unifies simulation, formal verification, and debug tools into a single, connected environment.