NEURA Robotics and the Technical University of Munich (TUM) has opened the TUM RoboGym powered by NEURA at LabCampus near Munich Airport (Germany). Spanning more than 2,300 square meters, the new center brings together research, robot training and industrial application. Robotics is increasingly being applied in healthcare, including in patient care and medical settings.
It was built in partnership with TUM’s Munich Institute of Robotics and Machine Intelligence (MIRMI). To mark the opening, NEURA is also launching 4NE1 Gen 4 (“For Anyone”), the new generation of its humanoid robot.
The RoboGym brings together AI research and industrial application. Companies contribute concrete tasks and their process knowledge. Researchers and development teams investigate how robots can learn these tasks faster and perform them more reliably.
Our companies hold knowledge that has been built up over generations. People know how to work materials, assemble components and master difficult processes. This knowledge is one of our greatest strengths in the age of Physical AI. With TUM, we are building a powerhouse for physical capabilities, where this knowledge becomes the foundation for new possibilities. The Mittelstand, in particular, should be able to help shape this future. With the launch of the humanoid 4NE1 Gen 4, we are showing today which robots are ready for it: designed for series production, built for productive work and powered by Neuron AI, our own foundation model that learns from experience in the NEURA Gyms. Developed, built and trained in Germany, for the world.
David Reger, Founder and CEO of NEURA Robotics
Partners have been training concrete workflows at the TUM RoboGym since September. Companies from more than ten industries are involved, including automotive, pharmaceuticals, medical technology, packaging, logistics, agriculture and care. Demand is high: capacity in Munich is already in strong demand.
Tasks range from handling cardboard boxes and filling bottles to loading and unloading wooden components. Companies can test applications, identify training needs and assess the suitability of different robot systems before expanding use in their own operations.
Various form factors are available for this: humanoid robots, mobile assistance robots such as MiPA, and cognitive robot arms. What matters is the capability a given task requires. To support this, the Center provides a shared development environment for academia, established companies, and startups.
The TUM RoboGym robotics center is unique in Germany and Europe. By collaborating with a robot manufacturer based in Germany, we are also sending a clear signal for Germany as an industrial location in geopolitically uncertain times.
Prof. Thomas Hofmann, President of the Technical University of Munich.
How does a robot grasp a delicate component? How does it recognize that two parts have been correctly joined? And how does it respond when a workflow changes? Capabilities like these are built through experience with the physical world. Physical AI requires data on how objects, materials and forces behave in the real world. The RoboGym combines three approaches to generate this data:
- Real-world robot data: Different robot systems perform tasks. Their sensors capture movements, contacts and the outcomes of their actions.
- Human demonstrations: People demonstrate workflows. Sensor-equipped data suits, motion-capture systems and teleoperation help translate movements and interactions into usable training data.
- Simulation and Data Amplifier: Virtual training environments expand real-world experience with additional variations and situations, allowing skills to be trained under changed conditions and then validated on real robots.
The Neuraverse connects this training infrastructure with NEURA’s software and AI tools. For the joint research conducted by TUM and NEURA, training data is structured and processed for use in developing and studying Physical AI models.
At the center of it all is a practical question: how quickly can a robot learn a new task, and how reliably can it be performed?
With 4NE1 Gen 4, NEURA is launching the new generation of its humanoid robot, already in operation at the TUM RoboGym for the opening: slimmer, lighter and powered by Neuron AI as its brain.
- New design, lighter body: 4NE1 Gen 4 has a slimmer body, a textile outer skin and is 20 percent lighter than the previous generation. The fabric makes the robot appear more approachable: it is built to work alongside people, and designed to look the part. The material also offers potential for further, simple customization.
- More precise handling: Advanced hands enable even more accurate grips. Added to this are Smart Limbs: reflexes are built directly into the limbs, so each limb can respond independently. This allows 4NE1 Gen 4 to react instantly to its environment and move more fluidly and in a more human-like way.
- Neuron AI, the brain: 4NE1 Gen 4 runs on Neuron AI, NEURA’s Perception-Language-Action model for Physical AI. It translates what the robot perceives directly into movement. Alongside vision, touch, force, sound and the robot’s own state also feed into the model. This makes a real difference for industrial tasks in particular, for example when a robot needs to recognize whether a fastener has clicked into place. Neuron AI is trained on real-world robot data from the NEURA Gyms, human demonstrations and simulation. Through the Neuraverse, skills learned by one robot become available to other connected robots as well. As a result, 4NE1 Gen 4 benefits from everything learned across the growing network of NEURA Gyms.
4NE1 Gen 4 is being prepared for series production. Alongside the opening of the TUM RoboGym in Munich, it is also appearing at the NVIDIA GTC conference in Berlin and is already in operation at the NEURA Gyms. At these training facilities, companies can train tasks with the new robot generation, making them among the first to work with 4NE1 Gen 4.
The TUM RoboGym gives researchers, students and companies access to a shared research and training infrastructure. Scientific findings can be tested here under practical conditions, while industrial tasks in turn generate new research questions.
NEURA contributes robotics, technical infrastructure, computing power and operating staff, among other things. TUM provides space, scientific expertise and resources for research and operations. The joint initiative is initially set up for three years. In total, the partners are investing around EUR 17 million, of which NEURA is contributing approximately EUR 11 million.
TUM MIRMI and NEURA Robotics have found two partners that can achieve great things together. Building on our world-leading scientific expertise, we are now able, together with a strong industry partner, to prepare concrete applications for the market and make industrial robotics an important factor in the German economy.
Prof. Lorenzo Masia, Executive Director of TUM MIRMI and Head of the TUM RoboGym.
The TUM RoboGym marks the start of an international network of NEURA Gyms. Further locations are being established in Europe, the United States and China. In Germany, another Gym is being built in cooperation with RWTH Aachen.
“We want companies to carry their experience into the next industrial era,” said David Reger. “To develop capabilities, protect their know-how and decide for themselves, in the future, where and under what conditions they make it available to others. What we are opening today in Munich is meant to have an impact far beyond this location. From Europe, together for the world.”
