The model is designed to control various types of robots and is explicitly aimed at industrial applications in the automotive, electronics, pharmaceutical, and logistics sectors. Particularly relevant is the approach of capturing real-world motion data from industrial workers via sensor-equipped gloves and using it to train transferable skills for robots.

For tasks where traditional automation often fails

Genesis aims to address precisely those tasks where traditional automation often fails: flexible assembly, wiring harnesses, sensitive components, and variable gripping and handling processes. The message here is clear: if robots no longer need to be rigidly programmed for every single task, but instead generalize skills from data and simulation, the scope of automation expands significantly. This also brings processes into focus that have previously remained the domain of humans due to product variety, cycle time uncertainty, or fine motor requirements.

New Hands on Robot Arms Instead of Humanoids

The market is moving in parallel on the hardware side. Linkerbot from China, a provider of highly mobile robotic hands with many degrees of freedom, is aiming for a valuation of $6 billion in its next funding round. The company reports holding more than 80 percent of the global market for robotic hands with many degrees of freedom and plans to expand production from around 5,000 to 10,000 units per month. Strategically, what is interesting here is less the humanoid show effect than the pragmatic industrialization: Linkerbot emphasizes that many customers mount the hands on existing robotic arms rather than purchasing complete humanoids.

Physical AI creates a new value chain

Established suppliers are also positioning themselves. Schaeffler expects a humanoid robotics order book in the triple-digit million-euro range by 2030 and, according to the CEO, is collaborating with around 45 robotics players worldwide. The group points to initial relevant contracts for actuators and components such as shaft gears. For the industry, this is an important signal: Physical AI is not just a software issue, but creates a new value chain comprising grippers, actuators, sensor technology, drives, gearboxes, training data, edge computing, and integration expertise.

AI as the operational core of the trend

At HANNOVER MESSE 2026, this shift toward AI-driven production was particularly evident. NVIDIA spoke there of a turning point for manufacturing and, together with partners, demonstrated applications involving AI Physics, agents, real-time simulation, vision AI, and humanoid robots in factory operations. Together with Microsoft, KUKA demonstrated how mobile robots and cobots can more intelligently orchestrate material flow, handling, and processing in a cloud- and data-integrated production environment. This is the operational core of the trend: AI is not introduced as a separate tool, but rather as a control and optimization layer overlaid on engineering, simulation, robotics, and the shop floor.

Practical knowledge from industrial applications

Research is also moving toward industrialization. At HANNOVER MESSE 2026, Fraunhofer IFAM presented the Machine Tool Robot, which is designed to bridge the gap between classic industrial robots and machine tools. Through model-based control, optimized drives, and mechanical structure, dynamic errors are to be compensated for and vibrations dampened, enabling higher path accuracy in demanding machining processes. At the same time, a training and competence center for humanoid robotics is being established in Stade, northern Germany, where practical knowledge from industrial applications will be captured, abstracted using AI, and transferred into system-independent models.

Not just a matter of efficiency, but a question of competitiveness

The geopolitical context heightens the relevance. The International Federation of Robotics reported on May 5 that China is placing robotics at the center of its modern industrial system in its 15th Five-Year Plan and is aligning AI research more closely with physical applications. China already has an operational fleet of around two million industrial robots; according to the IFR, China recently accounted for 54 percent of the industrial robots installed worldwide each year. For European companies, Physical AI is therefore not just an efficiency issue, but a matter of competitiveness.

New multi-layered safety mechanisms required

At the same time, a new category of risk is emerging. An article published on April 29, 2026, in Science Robotics warns that AI-enabled robots cannot be treated with the same safety mechanisms as chatbots. When foundation models control physical systems, traditional alignment approaches are insufficient; context-dependent, multi-layered safety mechanisms are required because malfunctions can trigger not only incorrect responses but also real-world damage.

Making existing production systems more flexible, robust, and scalable

While Physical AI is not yet a widespread replacement for traditional automation, it is the strategic litmus test for the next wave of automation. Companies should now identify which manual processes have not yet been automated—not due to a lack of cost-effectiveness, but because of variability, fine motor skills, or integration complexity. This is precisely where the new business case emerges: not in the humanoid robot as a symbol, but in adaptive, data-driven capabilities that make existing production systems more flexible, robust, and scalable.