Physical AI can Help India Move Beyond its Traditional Tech-Services Role

For India, the shift presents an opportunity to leverage its strengths in software and engineering while developing deeper capabilities in robotics, hardware, semiconductors and industrial technologies. The convergence of these disciplines will also require closer collaboration between industry and academia.

Against this backdrop, Pervinder Johar, CEO of Avathon, spoke to Deccan Chronicle on the evolution of Physical AI, India’s potential to become a developer rather than merely a consumer of these technologies, the talent and research ecosystem required, and the investments and policy support needed to build a globally competitive Physical AI ecosystem.

Physical AI is being described as the next major phase of AI adoption. What exactly distinguishes Physical AI from the current wave of generative AI, and where do you see its most immediate real-world applications?

The key difference is where the intelligence operates. Generative AI primarily works within the digital realm, helping people create, analyse and interact with information. Physical AI extends intelligence into real-world environments, enabling systems to interpret operational conditions, make decisions and support actions involving machines, assets and infrastructure.

Unlike conventional enterprise AI, which primarily operates within digital environments, Physical AI must account for the physical systems, environments and constraints in which AI is deployed. It is, therefore, inherently interdisciplinary, sitting at the intersection of AI, computer science, engineering and domain sciences.

Its near-term potential is particularly strong in manufacturing, logistics, energy and infrastructure, where organisations operate in complex and constantly changing environments. Applications could range from predictive and autonomous maintenance to supply-chain optimisation, asset management, worker safety and more efficient industrial operations.

India has traditionally been seen as a global services and software talent hub. What would it take for India to move up the value chain and become a serious developer of Physical AI technologies, rather than primarily a user or service provider?

India already has many of the foundational strengths needed to become a serious developer of Physical AI. The country has a deep pool of software and engineering talent, a growing deep-tech startup ecosystem, strong academic institutions and an increasingly capable manufacturing base.

The next step is to bring these capabilities together with greater expertise in robotics, industrial systems, hardware and advanced manufacturing. Moving up the value chain will require sustained investment in deep-tech research and development (R&D), closer collaboration between industry and academia, and greater access to real-world environments where these technologies can be developed, tested and scaled.

Most importantly, India needs to move from adopting technologies developed elsewhere to building solutions around the complex physical-world challenges it understands uniquely well. By combining its strengths in software, engineering and domain expertise with a stronger focus on hardware and industrial innovation, India can develop Physical AI solutions that are relevant domestically and can also be deployed and scaled globally.

Physical AI requires a combination of AI, robotics, engineering, semiconductor and domain expertise. Does India currently have the talent and research ecosystem needed for this convergence, and where are the biggest gaps?

India has strong capabilities across many of these individual disciplines. The bigger opportunity lies in connecting them effectively. Physical AI is inherently multidisciplinary, sitting at the intersection of AI, computer science, engineering and domain sciences. Success requires AI researchers to work closely with robotics specialists, engineers, semiconductor experts and professionals who understand specific industrial environments.

This is also where multidisciplinary academic institutions can play an important role. IIT Roorkee, for example, brings together expertise across areas such as civil engineering, geophysics, Earth sciences, hydrology and other engineering disciplines. Such breadth creates opportunities to explore how AI can be applied to complex physical-world challenges, rather than approaching the technology purely from a computer science perspective.

For Indian industries, adopting AI in the physical world could involve significant investments in robotics, sensors, connectivity, computing infrastructure and data. What should companies prioritise today if they want to be ready for this transition?

Companies should focus first on building the right foundations rather than rushing into automation. That means creating reliable operational data, connecting critical assets, improving visibility across processes and ensuring that their technology infrastructure is capable of supporting AI-driven applications.

The next priority should be identifying specific business and operational challenges where AI can deliver measurable value whether that is improving efficiency or reducing downtime or strengthening safety or optimising resource utilisation. Starting with clearly defined use cases allows organisations to demonstrate value while building the capabilities needed for broader adoption.

With the right data, infrastructure and use cases in place, companies can progressively move from AI-assisted decision-making to more advanced automation and, ultimately, greater autonomy in their physical operations. The transition is, therefore, less about adopting robotics or AI in isolation and more about building an intelligent operational foundation that can support increasingly autonomous systems over time.