Arm launches Total Design for Physical AI with 80+ partners

Arm is expanding its Total Design ecosystem program to physical AI applications, assembling more than 80 companies across the autonomous systems technology stack to tackle integration challenges that have slowed deployment of intelligent machines in industrial settings.

The initiative, announced by Arm, brings together partners including AWS, NXP, Siemens, Hugging Face, and Unitree Robotics to create common foundations for systems that combine AI models, sensors, compute hardware, and actuators. The program targets industries from mining and agriculture to manufacturing and logistics—sectors that represent trillions in economic activity but have proven difficult to transform with autonomous technology.

Why it matters

Physical AI represents a $200 billion annual compute opportunity by the 2030s, according to Arm's estimates, but fragmentation across the technology stack has created barriers to scale. Without shared standards and integration frameworks, companies building autonomous vehicles, industrial robots, and other intelligent machines face lengthy development cycles and deployment risk. A coordinated ecosystem approach could accelerate time-to-market while reducing the technical complexity that has kept many physical AI applications in pilot phase.

Robotics Capability Framework introduced

As its first collaborative effort, the ecosystem is developing a Robotics Capability Framework—a standardized language for defining and comparing robotic system capabilities. The framework aims to connect real-world use cases with specific system requirements including latency, compute placement, memory constraints, and safety characteristics.

Arm's chief architect Richard Grisenthwaite outlined the need in a manifesto, noting that robotics lacks the equivalent of SAE Levels, which created shared vocabulary for driving automation. The framework's initial structure drew from feedback across the robotics ecosystem, and Arm is inviting broader industry participation from companies including Anaxi Labs, ANYbotics, Fourier, and Lenovo to shape its evolution.

Proven model from cloud AI

The physical AI program extends Arm's Total Design model, which previously focused on cloud AI infrastructure. That approach brought together partners across the technology stack to reduce integration risk and optimize workloads from concept to deployment.

In automotive applications, the collaborative model has already shown results. Arm worked with AWS, Google, HERE, RemotiveLabs, and Siemens on an integrated digital cockpit reference solution that enables developers to test and validate complex automotive software on Arm Zena CSS before silicon becomes available.

The physical AI expansion adds capabilities including virtual platforms and digital twins to enable earlier development and validation of autonomous systems. Partners span software stacks, AI models, sensors, and compute hardware, addressing the reality that no single company can solve the full integration challenge alone.

The details were first reported by Arm in its newsroom announcement.