日本語 
Cadence Design Systems has teamed up with Google Cloud to take its chip design stack deeper into AI. The focus is ChipStack AI Super-Agent, now integrated with Gemini, and built to automate large parts of design and verification.
This is not just about adding AI on top. The idea is to shift how chip design actually happens. Today, teams still spend a lot of time across digital design, testbench creation, verification planning, and debugging. It is heavy, iterative work. ChipStack tries to compress that by letting AI agents handle execution across these stages.
The system combines Cadence’s EDA tools with agent-based inference. There is also something they call ‘mental modeling’ under the hood, which is meant to improve how the AI understands context and generates outputs with better accuracy. The claim is up to 10x productivity gains, along with shorter development cycles and faster path to tape-out.
Also Read: Japan–U.S. Chip Consortium Opens Silicon Valley Base to Accelerate AI Semiconductor Innovation
Google Cloud plays the infrastructure role here. It provides the compute needed to run Gemini and the AI agents at scale, while also enabling a cloud-native setup where teams can deploy and run workflows without building everything from scratch.
Step back and this is part of a bigger shift in semiconductor design. The complexity is exploding, timelines are tight, and manual workflows are hitting limits. Cadence is betting that agent-driven systems become the new baseline, where engineers guide the process and AI handles the grind.
English