Particle.news
Download on the App Store

Cadence and Nvidia Pair Physics Simulation With AI to Speed Robot Training and Build Data‑Center Digital Twins

The move signals a turn toward simulation‑first, agent‑led engineering spanning robotics to AI infrastructure.

Overview

  • Cadence and Nvidia, which announced the expanded collaboration Wednesday at CadenceLIVE in Santa Clara, will link Cadence’s high‑fidelity physics engines with Nvidia’s Isaac and related robot‑training models to shrink the sim‑to‑real gap.
  • Beyond robotics, Cadence will use Nvidia’s Omniverse DSX blueprint to create digital twins of large AI data centers so teams can test layouts, power and cooling in software before building the facilities.
  • Cadence also introduced agent‑driven design tools that tie its ChipStack and new AgentStack platforms to Google’s Gemini models, with Nvidia adopting AgentStack inside its own engineering flows.
  • Neither company provided dates for when joint tools will ship, and the near‑term market response was modest, with Cadence shares up about 2.46% Wednesday and Nvidia slipping slightly Friday.
  • Analysts say Cadence is pushing past traditional chip design toward system‑level simulation and automation, a shift that could let engineers find gains like higher tokens‑per‑watt and reduce costly trial‑and‑error in real deployments.