During Nvidia’s GTC-Munich event, Cadence unveiled three major breakthroughs in chip design automation. The announcements included a dedicated sovereign AI cloud for advanced design, a sophisticated agentic AI Super Agent tailored for printed circuit board workflows, and a strategic joint venture with Toyota Europe leveraging Nvidia Blackwell GPUs for high-fidelity external simulations.

A Sovereign Data Center Tailored for Chip Engineering

To better serve its European clientele, Cadence established a sovereign AI data center in Munich. This facility powers Cadence OnCloud services alongside NVIDIA Vera CPU systems, the Cadence Reality Digital Twin Platform, and the NVIDIA Blackwell-powered Cadence Millennium M2000 AI Supercomputer. Cadence stands out as the pioneer Electronic Design Automation vendor to introduce a sovereign AI framework in Europe, allowing local enterprises to engineer semiconductors, advanced packaging, and AI infrastructure while strictly adhering to European data residency mandates.

The newly introduced cloud environment merges three distinct tiers of AI-powered engineering:

  • Agentic AI: Utilizing NVIDIA Nemotron open models, the Cadence AuraStack AI Super Agent streamlines electronic system layout operations like schematic capture, routing, component placement, and multi-objective optimization.
  • Computational software: EDA and SDA platforms leverage GPU-accelerated NVIDIA CUDA-X libraries, including cuDSS and Omniverse, to speed up thermal evaluations, power integrity checks, electromagnetic mapping, and circuit simulations.
  • Accelerated computing: Millennium M2000 hardware delivers a robust, scalable infrastructure running on NVIDIA Blackwell GPUs to support both the AuraStack Super Agent and core engineering tools.

Expanding AI Super Agent Capabilities to Nokia

By scaling its agentic AI architecture, Cadence now offers an exceptionally comprehensive portfolio designed to drastically cut down time-to-market and boost engineering output. Originally introduced as the ChipStack AI Super Agent for system-on-chip verification, the lineup has expanded into four distinct solutions, which include InnoStack AI for digital implementation, AuraStack AI for packaging and PCBs, and ViraStack AI for analog customization.

Telecommunications giant Nokia is integrating the Cadence AuraStack AI Super Agent into an on-premises, secure environment to drive large-scale electronic system designs. This implementation boosts key analysis workflows by a factor of five while ensuring strict control over proprietary intellectual property.

AI gives engineering teams the speed to explore more designs and the insight to make better tradeoffs, while keeping proprietary data and expertise under their control, noted Tim Costa, vice president and general manager of industrial and computational engineering at Nvidia. Nokia is bringing that capability inside its own environment, with Cadence agents built with Nvidia Nemotron open models and simulation.

This deployment establishes an innovative benchmark for enterprise engineering teams. Nokia utilizes AuraStack to automate tedious tasks, evaluate complex architectural tradeoffs, and execute multiphysics simulations across diverse electronic configurations. Thanks to the modularity of the Millennium platform, Nokia can effortlessly expand capacity and execute numerous engineering workloads concurrently.

Accelerating Automotive Aerodynamics with Toyota

Moving beyond traditional EDA boundaries, Cadence has expanded its expertise into mechanical, biological, and system-level engineering. The firm is currently partnering with Toyota Motor Europe to assess advanced external aeroacoustic and aerodynamic simulations on the Aygo X vehicle via Cadence Fidelity Computational Fluid Dynamics software. This initiative highlights how accelerated computing reduces turnaround times while generating rich datasets to train physics-driven AI models.

Antoine Delacroix, senior manager of Vehicle Performance Engineering R&D at Toyota Motor Europe, stated that their work with Cadence proves how scalable, high-fidelity CFD on Nvidia GPUs delivers significantly faster results than legacy CPU workflows without compromising precision.

This initiative further highlights the massive performance scaling unlocked by enterprise-grade GPUs compared to traditional CPUs or workstation-class hardware. Benchmarks for the Aygo X simulation indicate that an eight-GPU B200 setup achieves convergence roughly five times faster than an RTX PRO 6000 Blackwell Server Edition configuration, while a workstation setup requires 32 GPUs to match the speed of just eight B200 units.