NEURA Robotics and Italian edge-computing company SECO are partnering to develop and manufacture compute modules for cognitive robots including the 4NE1 humanoid, while building a wider European infrastructure for Physical AI.
Europe’s humanoid robotics race is increasingly becoming a race to build the infrastructure behind the robots.
NEURA Robotics has announced a strategic partnership with Italian technology company SECO focused on the design, engineering and manufacturing of computing hardware for its cognitive robots, including the 4NE1 humanoid.
The collaboration will also extend beyond robot hardware into real-world industrial training, semiconductor manufacturing and a new NEURA Gym planned for Italy.
Building the Computing Backbone of 4NE1
Modern humanoid robots require enormous amounts of computing power.
They must simultaneously process camera feeds, sensor information, joint positions, language instructions and physical interactions while making decisions quickly enough to safely operate around people.
NEURA’s approach distributes some of that intelligence throughout the robot rather than relying on a single centralized computer.
The company calls this its Brain + Nervous System architecture.
Through its Smart Limbs concept, sensing and computing are positioned closer to the joints and physical components where rapid decisions need to happen. This allows parts of the robot to react locally rather than waiting for every piece of information to travel through one central processing system.
SECO will help engineer and industrialize the electronic compute modules behind this architecture.
The modules will incorporate Qualcomm Dragonwing processors, designed to support on-device AI processing, low-latency control and power-efficient edge computing.
From Prototype Hardware to Series Production
The partnership is particularly important because it addresses a challenge that appears once robotics companies move beyond prototypes.
Building several advanced robots is very different from manufacturing thousands of them reliably.
Electronics must be standardized, components need dependable supply chains, and computing systems have to perform consistently across every machine leaving the production line.
SECO brings experience in embedded computing and industrial electronics manufacturing, while NEURA contributes its robotics platform and Physical AI architecture.
According to the companies, the compute hardware developed through the partnership will be manufactured in Europe and designed to support scalable production of NEURA’s cognitive robots.
That could eventually include much larger numbers of 4NE1 humanoids if NEURA succeeds in scaling deployments.
Italy Is Getting a NEURA Gym
The partnership is not limited to manufacturing electronics.
NEURA and SECO are also planning a NEURA Gym in Italy, which would become the company’s first robot training hub in Southern Europe.
Robot training facilities are becoming increasingly important across the Physical AI industry.
Unlike conventional software AI, robots must learn how actions translate into movements and physical consequences.
Picking up an object, inserting a component, pushing a cart or handling a tool involves variables such as friction, weight, force, positioning and unexpected movement.
Training hubs allow robots to repeatedly perform those tasks and collect real-world data that can be used to improve their AI systems.
The planned Italian facility gives NEURA another location for building those datasets while working directly with European industrial environments.
Semiconductor Factories Could Become Training Grounds
One particularly interesting part of the partnership involves semiconductor and electronics manufacturing.
NEURA and SECO plan to deploy robots in industrial settings and collect real-world production data that can be used to develop new Physical AI automation systems.
The goal is to teach intelligent machines increasingly complex production processes and then turn those learned capabilities into reusable automation skills.
Instead of developing a robot specifically for one factory and leaving its knowledge there, NEURA wants useful skills learned from individual deployments to eventually become reusable across other robots and environments.
That idea connects directly with the company’s broader Neuraverse strategy, where robotic capabilities can function more like software that is developed once and deployed repeatedly.
Europe Wants More of the Robotics Stack at Home
The partnership also has a strategic dimension.
Many of the world’s most advanced AI processors are developed outside Europe, while much of the current humanoid manufacturing boom is concentrated in China and the United States.
NEURA and SECO are trying to keep more of the engineering, manufacturing and industrial knowledge surrounding Physical AI inside Europe.
SECO’s compute modules will be engineered and produced in Europe, while NEURA contributes robotics development from Germany and the planned training infrastructure expands into Italy.
The processors themselves will still rely on Qualcomm technology, making the project less about complete technological independence and more about building a stronger European robotics supply chain around leading global components.
The Humanoid Race Is Moving Beneath the Surface
Humanoid robots are usually judged by what people can immediately see: walking speed, dexterity, lifting ability or impressive demonstrations.
But many of the technologies that will determine whether these machines scale are hidden inside them.
Compute modules, sensors, power electronics, manufacturing processes and training infrastructure may ultimately be just as important as the robot’s external design.
NEURA’s partnership with SECO reflects that transition.
Rather than announcing another demonstration, the companies are concentrating on the less visible infrastructure required to turn advanced Physical AI systems into products that can be manufactured and deployed repeatedly.
For Europe’s humanoid robotics industry, that could prove increasingly important.
The next stage of the robotics race will not only be about building intelligent machines.
It will be about building the industrial ecosystem capable of producing, training and improving them at scale.
