'Physical AI' as a turning point for robot use?

Physical AI marks a fundamental change in robotics, moving away from pure automation towards autonomous action in the real world. Across all sectors, a majority of managers believe that fundamental changes are possible: in the high-tech sector, 93% see physical AI as a "game changer", in warehousing and logistics the figure is 69% and in agriculture 59%. This view is shared by almost three quarters of managers in the USA and around two thirds in Europe and the Asia-Pacific region.

From experiment to added value

Physical AI is at a turning point: technological breakthroughs and economic developments are accelerating its use in the real world on a large scale. Advances in foundation models are giving robots the skills they need to act autonomously in complex environments. At the same time, simulation technologies are significantly shortening training cycles by enabling learning on a large scale.

AI, robotics and data are mutually reinforcing: real data generated in the field continuously improves the performance and generalization capability of the systems. Added to this are advances in edge computing and battery technologies, falling hardware costs, new business models such as Robotics as a Service (RaaS) and breakthroughs in connectivity, for example through private 5G networks and precise wireless positioning.

The optimism is correspondingly high: 60% of managers assume that physical AI will enable robotics applications that were previously not possible or economically viable. The use cases range from hazardous activities, micro-logistics, pick-and-place tasks and field inspections to industry-specific scenarios such as dynamic assembly in manufacturing, support in the public sector's health and care sector and damage assessment following disasters in the insurance industry.

Accelerator for reindustrialization and operational resilience

With increasing reindustrialization in Europe and the US, physical AI is becoming a key accelerator for this change. 43 percent of executives say that reshoring and reindustrialization are increasing their interest in physical AI to support scalable domestic production. Two-thirds of companies now give physical AI a high priority in their automation agenda for the next three to five years. More than half of respondents see autonomous mobile robots, industrial robotic arms and collaborative robots (cobots) as the fastest growing forms of robotics in this period, well ahead of humanoid robots.

Labor shortages are a key driver for investments in physical AI, even before pure labor costs. This factor is particularly pronounced in agriculture, retail, the high-tech industry, warehousing and logistics and the automotive sector.

Physical AI also supports the necessary agility for long-term reindustrialization. Almost half of managers cite greater flexibility as a key benefit, particularly through the ability to reconfigure production systems and workflows more quickly than is possible with traditional robotics or rigid automation. In addition, more than half highlight improvements in safety and reduced physical strain for employees.

"Robotics used to mean factory floors, protected environments and defined and repetitive tasks. Universal use was out of the question. Technological breakthroughs in artificial intelligence, computing power and mechanics are now marking a turning point. AI now enables systems not only to see the world, but also to understand it and to act adaptively and autonomously. Physical AI is finding its way in the human world and can interact with us naturally," explains Peter Fintl, Vice President Technology & Innovation at Capgemini Engineering. "Even if the world is currently looking at humanoid robots, the lever for beneficial use lies directly in traditional designs such as automated transport systems, robotic arms or collaborative semi-stationary robotic systems. For European and German industry in particular, it is now important to quickly master the leap from isolated demonstrators to broad productive application."

Scaling physical AI and humanoid robots

Almost two thirds of managers expect physical AI to make the leap from pilot projects to broader implementation within the next five years - even though only four percent currently state that they are already operating on a large scale. In fact, scaling remains a challenge for almost eight out of ten managers.

Short-term growth is being driven by established forms of robotics. Although humanoid robots are arousing great interest, they continue to face considerable challenges and are therefore considered a long-term option: 72% of managers cite technological immaturity such as a lack of reliability and dexterity as an obstacle, 63% cite high costs and 58% see training requirements as a hurdle. In addition, the return on investment of humanoid robots is currently not clear enough for more than six out of ten managers. Social acceptance is also a challenge. More than six out of ten managers see public resistance as a critical obstacle to the introduction of humanoid robots.

Methodology of the study

In January and February 2026, the Capgemini Research Institute conducted a global survey of 1678 executives from companies with an annual turnover of more than 1 billion US dollars. The survey covered 16 countries in North America, Europe and the Asia-Pacific region and 15 industries. For the aerospace and defense industries and for the public sector, the revenue threshold was 500 million US dollars. The respondents were executives at director level and above.