Deloitte has published a new paper outlining how Physical AI — the integration of artificial intelligence with physical systems such as robots and industrial machinery — is shifting from early experimentation to large-scale deployment, even as most organisations remain underprepared for the transition.

The paper, titled “Physical AI: The Moment of Acceleration,” draws on Deloitte’s State of AI in the Enterprise survey and highlights a sharp divergence between current adoption levels and near-term expectations across sectors including manufacturing, logistics, healthcare, and technology.

Adoption still nascent, but acceleration expected

Only 5 percent of firms report that Physical AI is transforming their industry today, compared with 45 percent for traditional AI and machine learning. However, 41 percent of business leaders expect Physical AI to have a transformational impact within three years. The share of firms with Physical AI extensively integrated into operations is currently just 3 percent — a figure Deloitte forecasts will reach 18 percent within two years.

Adoption is expected to be highest in consumer, life sciences, and healthcare sectors at 22 percent each, followed by technology, media, and telecommunications at 18 percent, and energy, resources, and industrials at 16 percent.

Industrial robotics leads the way

Industrial robotics remains Physical AI’s most mature proving ground. Over 500,000 industrial robots were deployed in 2024, with annual installations forecast to reach 700,000 by 2028. Collaborative robots accounted for nearly 65,000 installations in 2024 alone. A Citi GPS report cited in the paper estimates that approximately 405 million robots of all kinds are currently in production globally, a figure projected to reach 1.3 billion by 2035.

“Physical AI marks the moment when intelligence moves off the screen and into the real world, transforming factories into learning systems that sense, decide and improve continuously. Organisations that are acting now will shape the operating models, skills and standards that define industrial leadership for the next decade.” — Chris Lewin, AI and Data Capability Leader, Deloitte Asia Pacific

Barriers to adoption

The main obstacles identified include cost and resource requirements cited by 41 percent of respondents, challenges in identifying use cases at 36 percent, talent and skills gaps at 33 percent, and technology or data availability at 31 percent. Deloitte emphasises that Physical AI implementation is as much an organisational challenge as a technology one — success requires aligning operational maturity with technology readiness.

Deloitte launches Physical AI Centre of Excellence in Shanghai

To support enterprises in scaling beyond pilots, Deloitte China has established a Physical AI Centre of Excellence in Shanghai. The centre combines Deloitte’s expertise in advanced manufacturing, robotics, and AI with immersive digital simulation technologies to help clients design, simulate, and deploy Physical AI solutions in real-world environments.

The full paper is available at deloitte.com.