At The Economic Times Future of Knowledge Work Summit in New Delhi, Adani Renewables Chief Digital Officer Kiran Nair made a firm case for sovereign AI built in India, for India. Speaking at a session on agentic systems and the future of work, he argued that national control over core AI infrastructure is essential as automation spreads across sectors.
Nair framed the issue in practical terms: building AI at scale demands compute, talent, energy, and data. He said India’s size, its large base of internet users, and diverse datasets give the country a head start—if it invests locally and sets clear rules.
Why Sovereign AI Now
Nair’s comments come as governments weigh how to secure critical technology and data. The term “sovereign AI” refers to domestic capability across training, deployment, and governance of models. For India, which has more than 800 million internet users by recent government estimates, the argument is also about shaping services for local languages and conditions.
India has poured resources into digital public goods like Aadhaar and UPI over the past decade. That effort fueled fintech growth and expanded access to services. Supporters of sovereign AI see a similar chance to set standards, reduce reliance on overseas providers, and protect sensitive datasets in health, finance, and public services.
“AI models need compute, talent, energy and data,” Nair said. “India’s scale, internet users and diverse data make local AI infrastructure critical.”
Four Pillars: Compute, Talent, Energy, Data
Nair outlined a basic checklist for national AI capacity. Each element brings its own challenges and openings:
- Compute: High-performance chips and data centers are essential for training and inference.
- Talent: Engineers, product leaders, and domain experts must work together on applied AI.
- Energy: Stable, low-cost, and cleaner power is needed to run large data facilities.
- Data: High-quality, diverse, and responsibly governed datasets drive better models.
India’s data center build-out is accelerating, with operators adding capacity near major metros. Energy is the pressure point. AI workloads are power-hungry, and grid costs can rise with demand. Here, Nair’s role at a renewable energy firm is telling. He pointed to green power as a way to support expanded compute while meeting climate targets and managing costs over time.
Agentic AI and the Workforce
The session’s title—“Agentic AI: The 10% Workforce That Will Replace 50% of Tasks”—signals a shift in how work gets done. Rather than replacing whole jobs, AI systems can take over repeatable steps, from document handling to forecasting. That can lift productivity and free up staff for higher-value work.
But the gains are uneven. Sectors with clear, rules-based processes will feel change first. Finance operations, customer support, logistics, and parts of public administration are early candidates. Nair emphasized the need to invest in skills so workers can supervise, audit, and improve AI agents.
“Local AI infrastructure” is not just hardware, he suggested. It is also people who can design, train, and validate systems in Indian languages and contexts.
Risks, Governance, and Balance
Building sovereign AI will test policy and industry. India still relies on imported chips, and supply chains are tight. Data protection and model transparency rules must keep pace. Clear procurement standards for government AI systems can set the tone for private adopters.
Experts also warn against fragmentation. If every firm or state builds closed stacks, costs rise and innovation slows. The middle path is to develop strong local capacity while using open standards and interoperable tools. Public datasets, privacy safeguards, and audit trails can help build trust.
What to Watch
Several signposts will show if India’s push is working:
- New data centers tied to renewable power purchase agreements.
- Investments to secure AI chips and shared compute for startups and research labs.
- Large language models trained on Indian languages and government datasets, with published safety tests.
- Upskilling programs focused on AI oversight, data curation, and product integration.
Nair’s case for sovereign AI is direct: India has the users, the data, and a motive to build at home. The next phase depends on aligning power, compute, and people with smart rules. If that happens, AI agents could handle routine tasks at scale while human workers move to higher-impact roles. The balance India strikes—between control and openness, speed and safety—will shape how quickly those gains arrive and who benefits.
