Arm launches Total Design for Physical AI and robotics framework

Details

Arm has launched Total Design for Physical AI, extending its collaborative development program (previously used for cloud AI) into physical automation. The initiative brings together more than 80 partner organizations — including AWS, Hugging Face, Liquid AI, NXP, Siemens, Unitree Robotics, and Qwen — behind standardized hardware and software baselines for robotic and physical AI systems targeting a 00 billion annual compute opportunity by the 2030s. Central to the launch is the Robotics Capability Framework, authored by Arm chief architect Richard Grisenthwaite. Modeled after the SAE Levels for driving automation, it classifies robotic systems across tiers of operational sophistication from reactive setups through context-aware, cognitive, and self-improving systems. Each tier specifies concrete parameters: latency, compute placement, memory allocation, power constraints, determinism, and safety standards. The practical pitch to hardware makers and software developers is reduced integration risk: standardized baselines mean less bespoke engineering when moving from proof-of-concept to production. Virtual platforms and digital twins are also included to accelerate pre-silicon development for automotive physical AI deployments. Whether the framework gains broad adoption will depend on how quickly partners actually build to it. The ecosystem breadth at launch — chipmakers, software firms, robotics companies, and cloud providers — is genuine, but framework announcements have an uneven track record in becoming industry standards without regulatory or market pressure behind them.