New Solutions for Physical AI

As the first in a series of basic robotics solutions from NXP, these ready-to-use solutions have been developed in collaboration with Nvidia. They implement the Nvidia Holoscan Sensor Bridge in conjunction with NXP's highly integrated SoCs. This reduces the number of discrete components, thereby reducing space requirements, energy consumption and costs. At the same time, the software complexity of robotic solutions for sensors and actuators is simplified, which also benefits humanoid robots.

Physical AI is considered the next level of innovation. It comprises systems that can detect, interpret and interact with their environment precisely, reliably and safely. The most sophisticated forms of physical AI include humanoid robots. To enable synchronized movements, dense sensor fusion and advanced actuator technology, they require secure, reliable and low-latency data processing and transmission across the entire robotic setup.

NXP's new integrated solutions for the robot body address these requirements directly. They deliver powerful edge intelligence and low-latency networking to enable safe and secure real-time communication. The solutions seamlessly integrate the Nvidia Holoscan Sensor Bridge into NXP's software enablement. This allows developers to easily implement real-time processing and establish a direct data connection between the robot body and predefined areas of the robot brain, significantly reducing latency. This makes it easier to integrate AI into physical systems, especially where real-time decision making is required.

"Physical AI is redefining what machines can do in the real world. Humanoid robots are the most complex manifestation of this evolution," says Charles Dachs, Executive Vice President and General Manager, Secure Connected Edge at NXP Semiconductors. "By combining NXP's extensive expertise in edge processing, secure networking, functional safety and real-time control with Nvidia's robotics platforms, we are greatly simplifying the development of physical AI. At the same time, we enable seamless connectivity between the Physical AI Edge and the central processing unit. This is just the beginning of what NXP will deliver to further accelerate the Physical AI ecosystem."

"The development of autonomous machines requires a high-performance computing architecture that can synchronize complex motor controls with perception in real time," says Deepu Talla, Vice President of Robotics and Edge AI at Nvidia.

The collaboration between NXP and Nvidia is helping to define a unified architecture for humanoid robotics with full-body control. NXP's edge processors, motor control MCUs and automotive-grade networking technologies form the basis of the system architecture. Added to this is the capability for asymmetric high-throughput data transmission acquired through Aviva Links. This is complemented by expertise in functional safety based on decades of automotive experience. Together with Nvidia's AI infrastructure, this creates a flexible and energy-efficient system architecture for next-generation robotic systems.

One of the first Holoscan Sensor Bridge-enabled solutions in NXP's robotics portfolio is a machine vision solution based on the i.MX 95 application processor. It transmits high-bandwidth data to the robot's central processing unit. The portfolio also includes a motor control solution based on a kinematic chain of i.MX RT1180 crossover MCUs and bundled via the S32J TSN switch from NXP. This establishes the direct connection to the central processing unit. The motor control solution supports integrated functions for common industrial protocols such as Ethercat and TSN.

The solutions will be available in the first half of 2026.