A new partnership has been formed to develop a European champion in the field of humanoid robotics. NEURA Robotics, a German startup founded in 2019 in Metzingen, south of Stuttgart, is the only German company focusing on humanoid robots. Their flagship product is the 4NE-1, a robot that can see, hear, and touch, powered by a "cognitive" technology developed by the company. The founder, David Reger, envisions robotics becoming as essential as smartphones. To achieve this, NEURA is moving toward mass production and has partnered with Seco, an Italian company specializing in embedded computing. Seco will design and manufacture the electronic modules that power the robot in Europe. These modules are central to NEURA's architecture, which distributes intelligence directly into the robot's limbs, allowing them to react in real time without waiting for commands from a central processor. NEURA and Seco also plan to use data from real production lines to develop automation solutions for the semiconductor industry. Their goal is to turn this industrial expertise into reusable building blocks that can be used beyond a single factory. However, the most strategic component of the 4NE-1, the Dragonwing processors designed by Qualcomm in the United States, is not European. According to Reger, Europe should approach physical AI by combining industrial expertise, developing the technology together, and deploying it globally. The partnership aims to scale production significantly, with NEURA promising millions of robots by 2030. However, neither company has provided specific numbers, production schedules, or volumes. The distributed architecture places part of the computation, sensors, and control directly into the robot's subsystems, reducing latency and improving robustness. However, this approach complicates integration, requiring synchronization of multiple controllers, data consistency, and secure internal communications. Embedded computing in a humanoid robot involves boards and processors that execute real-time perception, motor control, and local AI functions. An embedded module typically includes a processor, AI accelerator, memory, power supply, sensor/actuator interfaces, and industrial communication buses. It also includes security mechanisms such as watchdogs and secure boot. Modularity facilitates maintenance and industrialization but imposes strict constraints on thermal dissipation, power consumption, and reliability. The choice of an American processor, such as Qualcomm's Dragonwing, is strategic for sovereignty and the supply chain. The processor determines a large part of the performance, but also access to the supplier's drivers, development tools, and software libraries. Depending on an extra-European actor can create risks related to availability and technological control. This dependency influences the cost, requalification delays, and the ability to change platforms without redeveloping part of the embedded software.