Cadence, Dassault, Siemens, and Synopsys adopt NVIDIA NemoClaw to ship autonomous AI engineers

At its GTC Taipei keynote at COMPUTEX, NVIDIA showcased more than a dozen industrial-software companies — including Cadence, Dassault Systèmes, Siemens and Synopsys — building autonomous AI engineers on NVIDIA's NemoClaw blueprint. The agents are designed to automate the end-to-end workflow surrounding computer-aided engineering (CAE) and electronic design automation (EDA), compressing weeks of design, meshing, simulation, debug and reporting into hours.

What's new

• Cadence is building an autonomous register-transfer level (RTL) engineer with NemoClaw that orchestrates Cadence Design Systems ChipStack for digital design and verification — featured in a GTC Taipei keynote demo and described as cutting RTL verification from weeks to hours.
• Dassault Systèmes is productizing its 3DEXPERIENCE Agentic Platform on NemoClaw and NVIDIA OpenShell to operate long-running autonomous agents across design, simulation and manufacturing.
• Siemens is integrating NemoClaw and OpenShell into its Fuse EDA AI Agent, a purpose-built autonomous agent that plans and orchestrates multi-tool workflows across semiconductor, 3D-IC and PCB design.
• Synopsys is collaborating with NVIDIA on end-to-end engineering workflows with NemoClaw; Ansys Icepak (part of the Synopsys portfolio) is being demoed at COMPUTEX inside a NemoClaw-based agent for GPU electronics cooling.
• Startups building NemoClaw-based AI engineers include Flexcompute (multiphysics co-packaged optics), Luminary (long-running AI engineer for physics-model training loops), Neural Concept (electric motor design), nTop (geometry workflows used in JetZero's blended-wing-body program), PhysicsX (partnering with Microsoft Surface on a thermal simulation agent), P-1 AI ("Archie," a mechanical and electrical engineer), SimScale (cross-industry simulation agents) and Synera (injection-molding agent with Autodesk Moldflow).

Context

NemoClaw itself isn't brand new — NVIDIA has been positioning it for several months as an open blueprint for specialized, long-running agents with a secure runtime and frontier models. What's new is the breadth of vendor commitment in one keynote. NemoClaw includes a choice of harness, allowing integration with orchestration frameworks like NVIDIA OpenClaw and Hermes, plus a model router and NVIDIA NeMo libraries for customization. NVIDIA OpenShell, the open-source runtime at NemoClaw's core, governs how each agent accesses files, networks and tools and enforces policy-based security. Deployment targets include NVIDIA DGX Spark personal AI supercomputers, enterprise data centers and cloud service providers.

Why it matters

CAE and EDA are among the highest-value, slowest-moving enterprise software markets in technology — Cadence, Synopsys, Siemens and Dassault between them sit at the center of nearly every semiconductor, automotive and aerospace design pipeline on earth. Their simultaneous productization of NemoClaw-based agents is the strongest signal yet that long-running enterprise agents are the next major commercialization wave for AI in industry, distinct from chatbot-style copilots. It also tightens NVIDIA's platform position considerably: the agents run on NVIDIA-designed silicon, are orchestrated by NVIDIA's runtime, and are customized via NVIDIA's NeMo libraries. For competitors trying to break into the agent-runtime layer — whether AWS, Microsoft or specialized startups — NVIDIA's existing relationships with industrial-software giants are now a meaningful moat.