Siemens Uses Agentic AI to Speed Chip Design
Siemens announced it is accelerating integrated circuit design and verification by integrating agentic AI into its Questa One software. The AI-driven workflows are intended to speed up the register-transfer level (RTL) sign-off process for semiconductors.
The new Questa One Agentic Toolkit is composed of five distinct AI agents designed to automate specific pain points in the chip verification process. These include an RTL Code Agent for generating synthesizable code, a Lint Agent for checking design errors, a CDC Agent for clock domain crossing verification, a Verification Planning Agent for creating automated verification plans, and a Debug Agent for failure analysis. This toolkit is built upon Siemens' broader smart verification platform which has demonstrated significant performance gains. The platform has achieved up to 50x faster coverage closure, 8x faster gate-level design for test (DFT) simulations, and 48x faster fault simulation in specific applications. To power these agentic workflows, Siemens is leveraging NVIDIA Llama Nemotron models and NVIDIA NIM microservices. The toolkit is designed to be agnostic, integrating with mainstream AI coding applications like GitHub Copilot and Claude Code. The system is currently available through an early access program. Initial feedback from users like MediaTek has highlighted "immediate and significant" productivity gains, with engineers becoming proficient within hours. Tsavorite Scalable Intelligence, another early user, reported that the Lint Agent automated issue resolution, enabling a quick adoption of formal property verification. This move toward agentic AI addresses a critical industry challenge: the widening gap between escalating chip complexity and the availability of skilled verification engineers. As designs incorporate billions of transistors and complex 3D-IC and chiplet-based architectures, traditional verification methods have become a major bottleneck, consuming the majority of resources and time. By automating and optimizing verification, Siemens aims to shift the process from a reactive to an intelligent, self-optimizing system. The goal is to free up engineers from routine tasks to focus on higher-level architectural and system-level challenges, ultimately accelerating innovation and time-to-market.
