Arm and ST Advance Physical AI with Cortex-M

According to Arm and STMicroelectronics, the companies are expanding their long-standing collaboration to advance next-generation Arm Cortex-M microcontrollers for physical AI applications. At last week's STM32 Summit, STMicroelectronics announced a new strategic direction for its STM32 platform and introduced the STM32A series, a new family of Cortex-M-based microcontrollers designed to support safe and reliable autonomous systems.

As artificial intelligence (AI) moves beyond data centers and into industrial machinery, robotics, and vehicles, physical AI systems must be able to sense their surroundings, make decisions, coordinate movements, and respond in real time. These applications require precise and predictable processing while meeting strict power, size, thermal, and cost requirements.

Arm identifies three key computing layers for physical AI: perception-driven intelligence, which enables environmental awareness and autonomous navigation; interaction-driven intelligence, which allows machines to understand and respond to human instructions; and execution control, which manages motors, joints, and other actuators to perform precise physical movements. Cloud computing provides additional support for device management, simulation, and AI training, while functional safety and cybersecurity remain essential across all layers.

Arm and STMicroelectronics are working to advance the Cortex-M roadmap to address the growing demand for intelligent sensing, processing, and real-time control at the edge. The newly announced STM32A series is designed for autonomous systems requiring predictable performance and redundancy to maintain safe and reliable operation, even when certain components fail.

Drew Henry, Executive Vice President of Physical AI at Arm, said the collaboration builds on more than two decades of partnership and aims to bring greater intelligence to machines, enabling them to perceive and interact with the physical world.

Arm's Cortex-M portfolio already powers billions of devices, including industrial controllers and automotive systems. Processors such as the Cortex-M55 and Cortex-M85 also bring advanced digital signal processing and machine learning capabilities to resource-constrained applications. As intelligence moves closer to sensors and actuators, Cortex-M technology provides the precise, predictable, and continuous processing needed to translate AI decisions into physical actions.

Arm estimates that physical AI could create a $200 billion annual computing market opportunity across industrial, automotive, and other sectors by the 2030s. Continued collaboration and advances in embedded computing are expected to play an important role in enabling safer, smarter, and more capable physical AI systems.