India’s development story has long been measured in concrete and steel. Highways stretch across States, metro networks cut through urban corridors and airports expand to accommodate a growing economy. Yet alongside these assets, another infrastructure revolution is quietly unfolding, largely hidden from view.

Inside facilities with rows of servers, a new digital backbone is taking shape. Data centres are emerging as invisible highways of the digital economy, storing data, powering algorithms and enabling cloud services enterprises rely on. Driven by policy incentives and rising demand, India’s data centre sector is seeing significant multi-billion-dollar investments. Hyperscale infrastructure is expanding rapidly, positioning the country as a key node in the global digital economy.

Yet this expansion raises a deeper question: is India not only building the backbone of the AI economy, but also capturing its full value?

Hardware sans ownership

India’s rise as a destination for digital infrastructure rests on strong fundamentals; competitive costs, a large technology workforce and a fast-growing digital market. These advantages make it an attractive base for global cloud and AI infrastructure. However, this surge contrasts with more modest progress in domestic AI innovation.

India accounts for 2-3 per cent of global AI patent filings, while the US and China together dominate the field. While India ranks among leading countries in AI research output, its share of highly cited work remains limited. Public spending on research and development, at 0.6-0.7 per cent of GDP, continues to lag leading innovation economies.

This is reflected in global benchmarks such as the Stanford AI Index, where India ranks behind the US and China across key dimensions including research, investment and compute capacity. The pattern is familiar: strength in scale and adoption, with constraints in capital and compute-intensive innovation. The implication is clear. India is becoming an important location for AI infrastructure, but is still building depth in creating the technologies that run on it.

The servers may increasingly sit in India. The most powerful models often do not. In economic terms, the pattern is familiar. The inputs of the digital economy — data, infrastructure and engineering talent — are increasingly abundant domestically. The highest-value outputs — models, platforms and intellectual property — are still largely developed elsewhere.

This gap is not only visible at the national level; it is equally evident within firms. Many organisations have access to advanced digital infrastructure but are still evolving their ability to translate it into productivity gains. Cloud adoption often takes the form of migrating legacy systems rather than redesigning workflows around real-time data and automation.

A common pattern is visible across sectors. In one mid-sized manufacturing firm that recently migrated its enterprise systems to the cloud, production dashboards update in real time across plants. Yet procurement approvals still move through multiple layers of manual sign-off, delaying decisions by days. The data moves instantly; the organisation evolves more gradually.

This underutilisation mirrors a broader national challenge. The issue is not only access to infrastructure, but the pace of building capability around it.

The talent paradox

India produces one of the largest pools of technology professionals. Yet the requirements of the AI era are different. Artificial intelligence depends on deep research capability, advanced mathematics and sustained experimentation. While India has scale, it continues to expand the pool of researchers working at the frontier of AI development. Much of the ecosystem remains oriented towards implementation over foundational innovation.

This creates a talent paradox: a vast workforce with strong execution capability, alongside a growing but still limited base engaged in building core technologies. The emerging need is for “T-shaped” talent; professionals with deep domain expertise combined with the ability to apply AI tools across functions. This will require continued evolution in both corporate training and academic ecosystems.

Even as infrastructure expands, enterprise adoption of AI is progressing unevenly. AI is firmly on boardroom agendas. Large enterprises are building internal capabilities and experimenting across functions. However, mid-sized firms remain more measured, reflecting cost, organisational readiness and uncertainty around returns. Many organisations are developing structured approaches to measuring returns on AI investments. As a result, experimentation often precedes large-scale deployment.

A deeper issue is structural. Many firms are moving from layering AI onto existing processes to redesigning operations around it. Evidence suggests a small group is realising significant gains, while many remain in earlier stages.

The startup divide

India’s startup ecosystem reflects a similar duality. A segment risks falling into ‘AI-washing’, where artificial intelligence is emphasised even when the underlying technology is limited. Some firms act largely as intermediaries, building interfaces on top of global AI models rather than developing original capabilities.

At the same time, a smaller group pursues deeper innovation. Startups such as Sarvam AI and Krutrim are building language models tailored to India’s linguistic diversity. Qure.ai is applying computer vision to healthcare diagnostics, while CropIn is using AI to improve agricultural outcomes.

These efforts suggest that India’s AI innovation may evolve through domain-specific solutions, complementing global developments with locally relevant applications.

India’s digital infrastructure expansion is real and significant. Robust data centres, connectivity and computing capacity are essential for participation in the AI economy. India’s progress so far reflects a clear policy push towards building digital public infrastructure and enabling large-scale private investment in data centres and connectivity. This has created a foundation few emerging economies have matched. The next phase will require an equally sustained focus on innovation capability and enterprise transformation.

Capturing the full value of artificial intelligence will depend on how this transition is managed. Policymakers will need to complement infrastructure incentives with support for research and indigenous model development. Enterprises will progressively move from experimentation to embedding AI in core operations, while investors support deep-technology ventures with longer horizons.

India already possesses key advantages; scale, talent and data. The opportunity now is to translate these strengths into sustained innovation.

If current trends continue, India may host a significant share of AI infrastructure. The larger opportunity lies in ensuring it also participates in creating the intelligence that runs on it. The real challenge lies beyond the data centres; in building the capabilities that determine how value is created and captured in the age of artificial intelligence.

Sondhi is former MD & CEO of Ashok Leyland and JCB India, and Chauhan is with the Policy Unit of Quality Council of India. Views are personal

Even as infrastructure expands, enterprise adoption of AI is progressing unevenly. AI is firmly on boardroom agendas

Megha_Singh_Chauhan2 Vipin_Sondhi BL18_Think_Server_racks_data bl18-think1-main-vipin-sondhi

Published on April 18, 2026