Mechanical engineering: Is humanoid robotics on the verge of a breakthrough?

Germany is a leader in humanoid robotics research with institutions like KIT, DLR, and the University of Bonn. However, commercialization lags behind the USA and China.

“Our market analysis ‘2025 Humanoid Robot Study’ shows that 45 percent of companies are from China, 27 percent from the USA - while Germany has only one active company,” explains Tobias Bock, managing partner at Nexery. Given this global market distribution, the urgent question arises of how Germany can find targeted ways to generate sufficient capital - especially for startups in the scaling phase.

“Startups need significantly more ‘patient’ capital for scaling,” warns Patrick Schwarzkopf, managing director of VDMA Robotics and Automation. He therefore calls for a national strategy for mass production: “Scale baby, scale!” Tax incentives, simplified approvals, and support programs like the WIN Initiative could accelerate scaling.

Germany has all the core technologies for developing humanoid robots. Nevertheless, there is a lack of demand, especially from SMEs. High costs and investment risks deter, as the economic benefit is not immediately apparent. In addition, mature models that can be flexibly integrated into work processes are lacking. SMEs need practical solutions with clear added value, but so far there are only a few established areas of application. “Research institutions must work together interdisciplinarily to integrate insights from robotics, AI, neuroscience, and human-machine interaction,” explains Bock. “Practical test fields can significantly accelerate technology transfer.”

Germany relies on specialized industrial robots

Bock suggests that companies should specifically invest in German pilot projects and use open innovation formats such as consortia or industry partnerships to share development efforts and pool knowledge.

Ultimately, to promote growth, clear economic incentives and easier access to robotics technologies for SMEs are needed. "SMEs are already benefiting from plug-and-play solutions," says Dr. Eric Maiser, Managing Director of VDMA Research and Innovation, FKM e.V., and Head of VDMA Competence Center Future Business. "Interchangeable AI modules enable flexible upgrades without complete investments." This lowers entry barriers and makes humanoid robotics economically attractive.

The democratization of robotics starts right here. "Humanoids will elevate human-robot interaction to a whole new, intuitive level," explains Schwarzkopf. The USA and China are investing heavily in humanoid robotics, while Germany relies on specialized industrial robots. Strict labor law regulations and skepticism towards new technologies complicate integration. "Targeted funding programs and simplified approval processes are essential," Schwarzkopf demands. "We are currently pooling a lot of energy and networking stakeholders," he describes the VDMA strategy. The goal is concrete action plans for competitive series production in Europe.

Humanoid robots in production: These are the challenges

So far, humanoids in Germany are only used for simple, repetitive tasks such as transport or filling. However, a significant development leap is currently taking place: they are increasingly able to take on complex, fine motor, and cognitively demanding tasks such as assembly in timed production processes. In the 'Autonomous Factory' cluster of the VDMA Startup Radar, it is evident that many solutions for production automation and collaborative robotics are currently emerging, explains Dr. Eric Maiser.

If these projects are successful, humanoid robots could take over up to 40 percent of tasks in standardized production environments within three to five years. However, challenges still exist in speed, cognitive flexibility, and interaction in unstructured environments, but progress is being made.

Bipedal humanoid robots have made great strides abroad in recent years. Hyundai uses them in South Korea for assembly work in the industry. Amazon uses them in the USA and China for transport and sorting tasks. And South Korean shipbuilding will have them perform welding work in the future. Japan and China rely on them in elderly care and hospitals. In Japan and the USA, they welcome and inform guests in hotels and shopping centers. Unusual areas of application in the USA and Japan include companionship for lonely people, where they engage in conversations, remind them of medication, and provide emotional support. In China, they even compete against each other in the 'Beijing Humanoid Robot Sports Competition' in gymnastics and soccer.

“AI must be explainable and controllable”

Reservations come from people who have little contact with robotics. They are not even aware that they have been supported by and with AI for a long time. Of course: “AI must be explainable and controllable,” emphasizes Maiser. “This is initially a contradiction because AI as a 'black box' is a non-deterministic system. Training tools and demonstrators - such as in the style of an 'assistant buddy' - are means to build trust.”

Therefore, the VDMA recommends transparent communication and the close integration of the workforce when introducing such systems. “It is crucial that robots are understood as tools - not as replacements, but as extensions of human capabilities,” Maiser continues.

AI literacy, that is, the understanding and ability to deal with AI, is a crucial skill for people to use AI consciously and competently. This includes understanding AI fundamentals, critical thinking, ethical considerations, and practical skills in applying AI tools. The EU AI Act emphasizes these necessities. But one thing is certain: “The future belongs to hybrid skills - technological basic education combined with social intelligence,” explains Maiser.

Because it is just as certain that the shortage of skilled workers and labor is massively endangering our growth potential. "In this respect, we would be well advised to finally shelve the previous job killer debate," notes Patrick Schwarzkopf. "AI-based robotics and automation can help relieve existing workers from repetitive tasks and fill gaps."

And Bock adds that technologies like digital twins can make a decisive contribution: "They enable autonomous simulation-based training, continuous optimization, and the development of a software-defined robotics ecosystem."

Explainable AI is a must in safety-critical applications

But this requires modularization and volume production of humanoid robots, which drastically reduce costs. "The 'As-a-Service' and 'Pay-per-Use' approach have the advantage of low investment costs," explains Maiser. "This also applies to 'Robot-as-a-Service' models. They can lower entry barriers."

But Germany is far from being able to flexibly use humanoid robots in production, because what, for example, "the automotive industry is currently doing are rather first steps, and that in the truest sense of the word," emphasizes Schwarzkopf.

In this context, the need for ethical standards and transparency in the development of AI technologies becomes particularly clear. "Explainable AI, or XAI, is a must in safety-critical applications," explains Maiser. "The VDMA trend radar emphasizes that traceability and transparency are essential not only for regulatory reasons but also for acceptance."

For this, it is essential to anchor 'ethics-by-design', that is, ethical principles already in the development phase through diverse training data and documented decision-making processes. This is definitely a weak point, as programs and algorithms are developed by humans.

Security, personal rights, and data protection must also be ensured. Automation should always be considered from the perspective of the involved human. "We speak of the 'human-centered approach', as described in the Good Work Charter of the European robotics industry," explains Schwarzkopf.

Why ethical considerations are particularly relevant

These ethical considerations are particularly relevant when looking at the dynamic development in the technology sector. Startups often quickly turn technological trends into business models, which is why targeted networking with mechanical engineering companies is crucial.

“The successful AI startups that emerged in 2016, the year our future study 'Machine Learning' was published, were the reason for founding the VDMA Startup-Machine,” explains Dr. Eric Maiser. It accelerates cooperation through structured startup scouting, matchmaking, and platforms like the Startup Radar Dashboard. “Innovation challenges address specific innovation needs, while member coaching supports successful cooperation,” Maiser continues.

To promote humanoid robotics, VDMA R+A and VDMA Future Business rely on their own forum. Because “maybe we will also have humanoid robots at the mechanical engineering summit, so that this technology can be 'grasped' in the truest sense of the word,” hopes Patrick Schwarzkopf.

These companies will assert themselves in the long term

Humanoid robots are essentially “software-defined systems.” The crucial differentiation is made by the software. Companies that develop robust, scalable, and adaptively learning systems will assert themselves in the long term.

“A specific project will start in the second quarter,” announces Schwarzkopf. The next study by VDMA Future Business is dedicated to humanoid robotics and outlines various future scenarios.

Regardless, Helix demonstrates how generative AI allows humanoid robots to learn new tasks in real-time. "It is a so-called Vision-Language-Action (VLA) model that combines visual perception, language understanding, and precise motor control in a single system," describes Maiser. This technology is set to revolutionize human-machine interaction and enable flexible automation.

Data-driven services, such as real-time process optimization or predictive maintenance, open up new business fields. AI-based platforms can make robots active participants in value creation networks. "Our study 'Machine Learning 2030' showed that data-driven platform ecosystems accelerate the rollout," says Maiser.

"Biologization of industry" opens up new possibilities

Service robots could create emotional touchpoints and relieve staff. These developments correspond with the VDMA's trend map on AI-supported chatbots. Lightweight construction, biocomposites, and optimized energy efficiency make humanoid robots a potential contribution to the climate strategy. They could reduce CO₂ and temporarily compensate for disruptions like strikes or logistical problems.

“The 'biologization of industry' and 'biointelligence' open up a whole cosmos of new possibilities - from exoskeletons to DNA data storage,” explains Maiser. Artificial muscles could replace hydraulic systems, artificial skin enables precise interaction. AI models like Helix or Nvidia Isaac transfer skills between robots, creating learning effects at the system level, a crucial lever for scaling.