Capgemini Research Institute’s new report, “Physical AI: Taking human-robot collaboration to the next level,” says physical AI is moving from a niche idea to a planning priority for large organizations, but most companies are still early in the adoption curve.
The study is based on a global survey of 1,678 senior executives across 15 industries, supported by interviews with industry and academic experts. It looks at how AI is changing robotics, what business value companies expect, and what is still blocking scale.
Physical AI is no longer being treated as a side experiment, Capgemini says 66% of organizations now rate it as a high priority in their automation plans for the next three to five years. That figure rises in some sectors, including high tech (80%), industrial manufacturing (79%), and automotive (79%).
The report shows a clear gap between interest and scale; while 79% of organizations are already exploring, piloting, deploying, or scaling physical AI, only 27% say they are actually deploying or scaling it today. Capgemini also says 65% expect to reach scale within five years, which suggests broad intent but limited maturity for now.
The report treats physical AI as a shift in robotics, not just another automation tool. Capgemini says it enables machines to perceive, reason, and act in the real world, moving robots away from fixed, pre-programmed tasks and toward systems that can generalize across tasks, adapt to changing environments, and make context-aware decisions.
Capgemini also argues that the business case is growing because organizations are facing pressure; in the report, executives most often pointed to labor shortages (74%) and rising labor costs (69%) as the main drivers of investment. Competitive pressure and safety expectations also matter, but the first two drivers stand out because they affect day-to-day operations.
The report says physical AI is attracting attention because executives see it as useful in areas where traditional robotics has been limited. It highlights that 67% of executives view physical AI as game-changing, and 64% believe it will become a critical driver of competitiveness in their industry.
Other findings
Capgemini’s report “Physical AI: Taking human-robot collaboration to the next level” says the strongest near-term use cases are in tasks that are repetitive, physically demanding, or risky. These include hazardous operations, micro-logistics, pick-and-place, field inspection, and dynamic assembly. The report also points to healthcare, eldercare, and insurance-related inspection work as areas where physical AI may matter.
The report data also shows why scaling is still difficult; it says 73% of executives cite technology-readiness challenges as a major barrier, 71% say reliability is still a concern, 62% say robots do not yet have the dexterity required, and 56% say data scarcity limits effectiveness. Cost remains a problem too: 63% cite high costs and unclear ROI as a critical barrier.
That helps explain why the market is still dominated by pilots, proofs of concept, and limited deployments rather than large rollouts. Capgemini’s data shows that many companies are engaged, but only a smaller share has moved beyond testing into operational scale. The report also suggests that humanoid robots are still a longer-term bet, not a short-term deployment answer.
[Related Post: HCLTech and NVIDIA Launch Physical AI Innovation Lab in Santa Clara ]
What is physical AI, and why is it gaining momentum?
Physical AI refers to AI systems embedded in robots or other machines that can sense their environment, interpret what is happening, decide what to do, and act without constant human instruction. In the report’s framing, this is different from traditional robotics, which usually follows a fixed script and works best in structured settings such as assembly lines or controlled warehouse aisles. Physical AI is meant to handle more variable environments.
Its momentum is being driven by several forces at once, Capgemini says advances in multimodal foundation models, better simulation, falling hardware costs, and the rise of robotics-as-a-service are all lowering adoption barriers. The report also points to aging workforces and persistent labor shortages as structural reasons companies are paying attention now.
The report’s larger conclusion is that physical AI is moving from theory to practice, but the transition is uneven. Interest is high, pilots are common, and the technology is improving quickly. Still, the report makes clear that reliability, safety, data, and cost will decide how fast physical AI moves beyond the pilot stage.
View or download Capgemini’s full report, “Taking human-robot collaboration to the next level“
