Siemens AI EDA Portfolio Automates 3D Chip Thermal Design
The integrated workflow automates 3D IC electrothermal analysis, promising to cut engineering effort by 95 percent.
Updated on Oct. 26, 2026 in Semiconductors
Live Poll
Do you believe that using AI to automate complex engineering tasks improves the quality of technology?
Siemens has expanded its agentic AI Electronic Design Automation (EDA) portfolio to automate 3D integrated circuit (IC) electrothermal codesign. This new workflow, which is now announced, integrates Siemens tools with NVIDIA accelerated computing to handle complex thermal analysis.
Why it matters
Rising power density in advanced 3D chip architectures makes early thermal analysis critical to design success. This automation aims to address these physical constraints by shortening development cycles and increasing the volume of design exploration.
The workflow iterates through 200-500+ design points per cycle in an 8 to 10-hour runtime. It leverages NVIDIA Rubin GPU inference capacity at 50 petaflops with 288 gigabytes of HBM4 memory per GPU to manage the computational load.
The players
Siemens
A multinational technology conglomerate providing industrial software, automation, and EDA tools for semiconductor design.
NVIDIA
A semiconductor company specializing in accelerated computing platforms and AI hardware infrastructure.
The details
Engineers set design objectives, and an AI agent autonomously orchestrates seven Siemens EDA tools—including Innovator3D IC, Solido Characterizer, mPower Digital, Calibre 3DThermal, Simcenter FloTHERM, Calibre xACT, and Solido SPICE. The system evaluates design alternatives against signoff-grade physics-based solvers to identify root causes in power delivery, die placement, and material properties. This loop enables the system to refine configurations by iterating through tradeoffs until targets are reached.
Timeline
June 2026: Siemens introduced agentic AI capabilities at DAC 2026.
October 26, 2026: Siemens announced the expanded agentic AI EDA portfolio.
The Tech Race
This release marks a shift toward full-stack autonomous EDA, moving beyond standalone tools toward orchestrated agentic workflows. It positions Siemens directly against growing industry efforts to solve the thermal density limits of 3D ICs through high-performance simulation.
Chip designers and engineering teams can now offload complex electrothermal analysis to automated agents, significantly reducing manual configuration time. This workflow is designed to scale with existing high-performance computing infrastructure using NVIDIA hardware.
The takeaway
The move to agentic EDA signals a new standard for design throughput, effectively replacing labor-intensive thermal modeling cycles. Watch for future benchmarks demonstrating how these automated solvers perform against legacy manual signoff processes in production silicon projects.
Further reading
For broader trends in chip design, see the latest updates in Semiconductors.
Source note: This article includes information reported by EEJournal.
Live Poll
Do you believe that using AI to automate complex engineering tasks improves the quality of technology?
