Existing industrial robot solutions are based on clearly defined and repeatable motion sequences that are designed by experts and programmed in advance. As a result, conventional automation reaches its limits when faced with high product variety, variable processes, and assembly tasks that require complex manipulation. Moreover, existing models are not well suited to dynamic, unstructured production environments and changing tasks. This makes new automation approaches necessary to overcome current technological limitations and strengthen the competitiveness of European industry.

AI-powered robots expand the potential of automation through advances in software and hardware

AI-powered robots, often referred to as next-gen robots, no longer depend solely on predefined programs. Enabled by physical AI, they can navigate dynamic environments, make decisions in real time, and perform tasks autonomously. As a result, humanoid robots, quadrupeds, and drones are increasingly able to take over activities that are currently performed manually. This opens up new opportunities to automate production processes that were previously considered difficult to automate, particularly those involving high product variety or complex manipulation tasks requiring a certain degree of dexterity. Against this background, AI-powered robots are well suited for tasks such as assembly, machine loading, remote maintenance, data collection in hard-to-reach or hazardous areas, as well as for automated quality and safety checks.

Many manufacturing companies are already piloting the use of bipedal and wheeled humanoids as well as quadrupeds. Despite these advances, many pilot projects still face significant challenges.

1. Limited task complexity: Current applications mainly focus on highly standardized and repetitive tasks. Robots only need to respond to environmental changes to a limited extent. In many cases, the required cycle times are not yet achieved.
2. Limited manipulation capabilities: The tasks performed so far typically require only minimal force, a low degree of dexterity, and relatively simple handling procedures. Applications are limited to grasping, transporting, and placing objects. More complex assembly or joining processes remain challenging.
3. Limited human-robot interaction: Many pilot projects are still implemented in restricted or clearly marked working areas. As a result, direct collaboration between humans and robots remains limited, particularly in dynamic production environments.

To fully leverage the potential of AI-powered robots, further advances in software, hardware, and safety are required. On the software side, robotic intelligence and the underlying foundation models must be developed further so that robots can reliably perceive, decide, and act in varying production environments. A key challenge is training foundation models with high-quality industrial data. Whereas foundation models are increasingly trained using imitation learning and simulation-based approaches, transferring these capabilities reliably to real production environments remains difficult. On the hardware side, further development is needed in high-performance batteries for mobile applications, multimodal sensor systems, force-sensitive actuators and gearboxes, as well as dexterous manipulation systems such as robotic hands. In addition, safety standards and certification frameworks must continue to evolve, particularly for applications involving human-robot interaction.

Investments and strategic partnerships are gaining importance across the value chain of AI-powered robotics

Investment in AI-powered robotics is currently increasing significantly. Capital is flowing into the production and scaling of humanoid robots, foundational AI models for autonomous robotic systems, and performance-critical components such as actuators, gearboxes, grippers, and robotic hands. Interest is no longer limited to individual robot manufacturers but increasingly extends across the entire value chain. As a result, a broader ecosystem is emerging that could accelerate the development, market readiness, and scaling of AI-powered robots.

In addition to investments, strategic partnerships are becoming increasingly important for the industrialization of AI-powered robots. Manufacturing companies can play a dual role in this context: as suppliers and as industrial users of robots. The collaboration between Schaeffler and NEURA Robotics illustrates what this can look like in practice. Schaeffler is involved in the development and supply of key components for humanoid robots, particularly innovative actuators, thereby contributing its own technology and manufacturing expertise to the robotics value chains.

At the same time, Schaeffler plans to deploy NEURA humanoids within its own global production network and integrate several thousand humanoid robots by 2035. With this approach, Schaeffler is pursuing two goals: as a supplier, the company supports the development of future-relevant AI-powered robots while also testing, refining, and scaling them within its own production environment. This example shows how manufacturing companies can use their existing technological expertise to enter the market early and explore new business fields. In doing so, they not only create the basis for new automation opportunities but can also participate in the emerging value chain of AI-powered robotics.

Three regions are playing a key role in the development of AI-powered robotics

As part of a structured desk study conducted by the Fraunhofer Institute for Production Technology IPT, more than 190 companies in the field of AI-powered robotics were identified and mapped along the value chain.

Three key groups of players can be distinguished: next-gen robot manufacturers, software developers, and component manufacturers. Next-gen robot manufacturers develop embodied robotic solutions such as humanoid robots, quadrupeds, autonomous mobile manipulators, or cobots. Depending on their internal capabilities, they typically develop components and/or foundation models either in-house or in collaboration with selected partners. Software developers provide the foundation for robotic intelligence, perception, motion planning, and manipulation. At the same time, several companies are developing solutions to make high-quality industrial data available for training and further improving AI-powered robots. Component manufacturers supply the hardware foundation, ranging from sensors and actuators to dexterous gripping systems.

A look at the global landscape reveals three regions with distinct focus areas:

Asia

Asia, and China in particular, has the highest concentration of AI-powered robot manufacturers, with Shenzhen standing out as a major robotics hub. China therefore plays a defining role in Asia’s AI-powered robotics landscape. However, many current solutions still face limitations in terms of precision and dexterity. While software developers and component manufacturers are also important parts of the ecosystem, they appear to be less prevalent than AI-powered robot manufacturers. Shenzhen’s strong robotics ecosystem supports rapid prototyping, development, and industrialization.

North America

North America is home to several renowned AI-powered robot manufacturers that have already successfully demonstrated pilot use cases in industrial environments. The region also stands out for its strong software expertise, with many companies developing robotic intelligence. North America’s strengths include distinctive capabilities in foundation models, such as vision-language-action models, robotics simulation for training and optimizing AI-powered robots, and an established venture capital infrastructure that accelerates the development of AI-powered roboticst.

Europe, including Israel

Europe, including Israel, appears to have a more balanced ecosystem in which all three groups of players are represented. The region can also build on an established industrial automation industry. Hardware expertise, software development, industrial integration, and real-world application are closely intertwined. Europe therefore has a strong foundation for further developing AI-powered robotics across the entire value chain. In addition, Germany’s large SME base can serve as a valuable testbed for piloting and further developing AI-powered robots, particularly in (flexible) production environments where certain processes remain difficult to automate.

Europe’s strength lies in its ecosystem of robotics, component, and software companies

In addition to robotics companies such as NEURA, Agile Robots, and Humanoid, established component manufacturers including Bosch, Schaeffler, and Maxon are increasingly positioning themselves as strategic hardware partners for AI-powered robots. At the same time, software ecosystems are growing, with companies such as Wandelbots, Sereact, and Genesis AI. Alongside physical AI, the development of vendor-independent software and robotic intelligence plays an important role. Research by Fraunhofer IPT also shows that several robotics hubs are emerging in Europe. Munich and Zurich in particular, as well as Odense and Paris, bring together research institutions, industrial companies, and start-up activity.

Why companies should start building capabilities in AI-powered robotics now

AI-powered robotics is currently characterized by a highly dynamic market environment and is increasingly becoming a strategically important field for manufacturing companies. Investments are being made across the entire value chain, from robot manufacturers and physical AI to performance-critical components. At the same time, manufacturing companies have the opportunity not only to use new automation solutions but also to position themselves as technology partners or suppliers in the emerging value chain of AI-powered robotics.

This creates three reasons why manufacturing companies should explore AI-powered robotics today:

1.  Leverage the untapped automation potential across production: AI-powered robots, such as humanoids or quadrupeds, enable the automation of tasks in unstructured, highly variable, and human-centered environments without the need to fundamentally redesign production processes or machinery..
   
2.  Understand market signals early and build integration capabilities: Companies should systematically monitor the market, relevant players, and technological maturity, make targeted use of research and development partnerships, and build the internal capabilities needed for technology selection, piloting, and seamless integration.
3. Position early in the emerging value chain of AI-powered robotics: Besides deploying AI-powered robots, companies can diversify their business and position themselves as a technology partner for AI-powered robotics, e.g., in components, software or system integration

Consortium study: The use of AI-powered robotics in industrial production

As part of the “Next-Gen Robotics” consortium study, Fraunhofer IPT invites companies to join its AI-powered robotics community of industrial partners and work with selected robotic experts to discuss and address relevant challenges. In addition to hands-on workshops, the consortium study focuses on identifying technologically feasible and economically viable applications, assessing business cases, and developing validated technology roadmaps to support smooth implementation.

As an alternative to the AI-powered robotics community and consortium study, companies can evaluate technologies and application scenarios within their own organizations in a hands-on format such as Fraunhofer IPT’s Experience Day. Together with robotics experts, participants can discuss specific application opportunities, integration challenges, and technological diversification potential. In doing so, companies benefit from Fraunhofer IPT’s extensive robotics network.

Disclaimer: The Fraunhofer IPT analysis of companies in the field of AI-enabled robotics includes only those companies that have either documented their ambitions in the field of physical AI through official press releases or published specific use cases for AI-enabled robots (e.g., humanoids). Furthermore, the current mappings may change with ongoing developments and future updates.