All the computing power of the humanoid robot is contracted by AMD

Zhixi Author | ZeR0 Editor | Mo Ying Zhixi reported on July 28 that recently, at the Advancing AI 2026 Conference (AAI 2026), AMD released the Kria AI solution, a software and hardware "family bucket" for physical AI, which plugs the heterogeneous computing capabilities of AMD CPU, GPU, NPU and FPGA into the same robot architecture. Physics AI is the next frontier for AI to leave the screen and enter the real world to take action. In the digital world, an agent plans and completes tasks; in the physical world, it faces more complex challenges, including sensing the environment and motion, navigating under uncertainty, controlling real actuators, and protecting the safety of those around it. AMD's adaptive computing technology has been supporting sensing, safety and real-time control systems for industrial and surgical robots for more than 20 years. Today, AMD has further launched the Kria AI solution, extending AMD's advantages in robot bodies to robot brains, integrating AI perception, reasoning, and intelligent agent decision-making and control capabilities on a single platform optimized for stringent power consumption, heat dissipation, and latency requirements. Kria AI solutions include Kria AI system modules (SOMs) and robot carrier boards powered by AMD Ryzen AI Embedded At this point, a complex humanoid robot can be equipped with AMD computing products in many places: Kira AI: the "brain" of the robot, responsible for advanced perception and reasoning. The second generation Versal AI Edge: the "spine" of the robot, responsible for high-speed data and adaptive computing. Zynq UltraScale+: Robot “joints” and real-time control. Spartan UltraScale+: Sensor Connectivity &