Socionext Inc. is expanding its use of agentic artificial intelligence (AI) across system-on-chip (SoC) development, extending AI-driven workflows from front-end design and verification to the physical design of large-scale chips manufactured using advanced process technologies. The initiative includes a collaboration with Cadence Design Systems Inc. and the deployment of its InnoStack AI Super Agent for physical design work on a large-scale 3-nm SoC. Socionext has been integrating AI agents into front-end design and verification to improve engineering productivity, support technical analysis, and assist decision-making. The company is extending this approach beyond individual task automation by incorporating AI agents into broader development workflows to improve design quality and let engineers focus on complex technical decisions and innovation. Physical design presents additional challenges as SoC complexity increases and advanced process nodes impose demanding requirements on implementation, timing, and optimization. Integrating AI agents with electronic design automation ( EDA ) tools and established design flows can help engineers manage these tasks more efficiently. Through its collaboration with Cadence, Socionext is applying AI-driven methods to the physical design of a large-scale 3-nm SoC using the InnoStack AI Super Agent. The company plans to extend this approach to additional SoC development projects. Socionext is combining retrieval-augmented generation (RAG), Model Context Protocol (MCP) support, customized AI skills, and AI agents to incorporate internal engineering knowledge into practical physical design workflows. Current and planned applications include floorplanning, customized timing-analysis checks, and automated checklists. These capabilities aim to make proprietary design expertise more accessible and reusable across engineering teams and projects, reducing reliance on knowledge held by individuals. Beyond automating engineering tasks, Socionext aims to preserve accumulated design knowledge as a reusable organizational resource. This approach is intended to support consistent engineering practices, maintain design quality, and facilitate knowledge transfer between teams and future generations of engineers. Looking ahead, Socionext plans to connect its existing front-end AI initiatives with its expanding physical design capabilities to establish a more integrated AI-driven SoC development flow. The company will continue collaborating with Cadence while developing custom AI agents and engineering assistants that leverage internal expertise to improve productivity and knowledge reuse across its design organization. The post Socionext, Cadence Advance AI-Driven 3-nm SoC Design appeared first on Embedded .