Reflecting on the findings from Capgemini’s recent report, Dr Diane Berry, Engineering Science Leader at Capgemini, explains the changes needed to enable the UK to adopt physical AI fully and confidently.
Some technologies explode onto the market, whereas others advance steadily as organisations, regulation, and operating models catch up. Physical artificial intelligence (AI), in the UK at least, firmly belongs in the latter category. Robotics has a long history of inflated expectations, which has contributed to understandable caution among many organisations when it comes to adoption.
But while adoption may be measured, the potential impact is anything but. By combining AI, robotics, and advanced sensing, physical AI is set to transform how industries operate – enabling machines to perceive, reason, and act autonomously in the physical world. From critical infrastructure and manufacturing to energy and logistics, it has the potential to reshape productivity, resilience and safety at scale.
Research tells us UK executives are ready to look beyond the hype and recognise that strategic importance. According to Capgemini’s recent report, around two thirds (67%) say physical AI will become a critical driver of competitiveness in their industry.
Still, that recognition does not yet equate to full scale commitment. While 65% of UK executives now rate physical AI as a high priority over the next three to five years, the path to scale remains deliberate. The UK’s approach is measured and incremental, particularly when compared with faster-moving markets such as Japan, South Korea, China, and the US, where physical AI is already embedded as a core pillar of industrial strategy.
In contrast, the UK is progressing through structured exploration and early deployment focused less on acceleration, and more on building confidence, trust, and long-term viability.
Building the foundations for scale
UK organisations are balancing opportunity with responsibility, particularly in safety-critical and regulated environments. Physical AI is not a plug and play technology: it requires deep integration across data, operations, infrastructure, and workforce design. That reality naturally favours a more staged adoption curve.
Legacy operating models continue to shape the pace and nature of change. Historically, large-scale transformation in the UK has often prioritised stability, efficiency, and risk reduction over rapid experimentation and iteration. While effective in the past, these models can make it harder to build the hands-on operational capability needed to scale emerging technologies such as physical AI at pace.
Without strong internal understanding and collaborative, long-term partnerships focused on building capability, scaling becomes harder. Physical AI demands continuity not just of technology, but of skills, governance, and operational ownership.
The need for long-term commitment
There is also a broader behavioural challenge around investment. Physical AI requires sustained commitment over multiple years. Yet, UK organisations have often been cautious investors quick to pause or withdraw funding when short-term pressures emerge.
This stop-start pattern is fundamentally misaligned with technologies that mature through iteration, learning, and cumulative capability building. One-to three-year planning cycles are unlikely to unlock the full value of physical AI. What is needed instead is a shift towards long-term capability investment, anchored in clear vision, stable funding, and an acceptance that ROI will evolve over time rather than arrive instantly.
Where physical AI delivers first
The opportunity for those prepared to stay the course is significant. Unlike agentic AI, which has entered organisations through high volume, lower risk digital tasks, physical AI is often deployed first where the stakes are highest. It is best suited to high-risk hazardous and safety-critical environments – the ‘un humanable’ tasks where continuity, precision, and risk reduction are paramount. In nuclear power plants, for example, robotics can already perform tasks that are too dangerous for humans, like decommissioning or remote handling in high-radiation zones.
In sectors such as energy, infrastructure, manufacturing, and logistics, physical AI is already improving resilience: reducing downtime, enhancing safety and providing a stable, 24/7 operational backbone for essential infrastructure at a time of persistent labour shortages.
From exploration to action
Encouragingly, the foundations are being laid. UK organisations in our report are actively engaging with physical AI, moving from exploration into pilots and early deployments. Nearly half (47%) expect humanoid robots to work alongside human employees by 2030, signalling growing confidence in human-robot collaboration, albeit on longer timelines than some global peers.
This is not about rushing towards humanoids, nor about replacement. It is about convergence. Robotics, automation, and advanced sensing are not new, what is changing is their intelligence. Physical AI is enabling these systems to learn, adapt, and operate with far greater autonomy, transforming robots from tools into collaborators.
If the conviction expressed by UK executives is any indication, the intent is there. The next step is translating that intent into consistent action. When that shift happens, the opportunity moves decisively beyond cautious exploration and towards a form of scale that is resilient, trusted, and globally competitive.
Please note, this article will also appear in the 26th edition of our quarterly publication.