Shanti Bill to power India’s AI and data centre push: Vaishnaw

As AI adoption expands, Vaishnaw said countries will need a large number of data centres that require round-the-clock, high-capacity power. The data centres that power AI are energy-intensive facilities that also consume large amounts of water for cooling.

Vaishnaw argued that nuclear energy is best suited to meet these needs, as it can provide a stable base load without carbon emissions, making it central to preparing India for an AI-driven economy under the Shanti Bill.

He also pointed to advances in nuclear technology, including small modular and micro reactors, which can generate anywhere between 15 and 30 MW of power and be deployed in compact, container-sized units. Such reactors, he said, could make it easier to supply clean energy even in remote regions like Ladakh, while safety standards continue to improve.

The AI ecosystem rests on three layers, he explained. “The top layer is the application layer. The middle layer is models – LLMs, SLMs (large and small language models), focused models. The bottom layer is the infra layer – chips and data centres. We are working on all three,” Vaishnaw said while interacting with the media on the sidelines of the announcement of projects approved under the Electronics Component Manufacturing Scheme (ECMS) in New Delhi.

Investment in AI and data infrastructure is also picking up pace. “All over the world, there was a stir around the Stargate project. In India, there is already a committed investment of $70 billion in the AI, data centre infrastructure. So that’s the pace at which things are happening,” he said.

The Stargate project was launched in early 2025, led by OpenAI, SoftBank, Oracle, and others, to build large-scale AI data centres and support infrastructure to train and run advanced AI systems. It is intended to invest up to about $500 billion over the next few years.

Global tech companies such as Microsoft, Amazon and others have announced large expansions of cloud and AI infrastructure in India, while Google is set to build its largest AI data centre campus in Visakhapatnam, Andhra Pradesh, through a $15 billion investment over five years in partnership with AdaniConneX.

The government is also working on a national data centre policy framework, which is still under discussion and not yet finalised, according to a draft policy framework seen by HT. The draft proposes steps to make it easier to set up data centres through single-window clearances, streamlined approvals, and dedicated building norms. It also focuses on ensuring reliable power supply, encouraging renewable energy use, improving telecom connectivity, and promoting water-efficient cooling systems.

The draft framework also looks at classifying data centres as essential services, creating data centre parks, and supporting skill development through specialised centres for AI, cloud computing and cybersecurity.

On semiconductors, Vaishnaw said India is on track to become a major global player over the next decade. “In 2022, the semiconductor policy [India Semiconductor Mission (ISM)] came. We made good progress in design. By 2032, we would be in a very significant- among the top four nations in semiconductor technology. And by 2035 we would be among the best. This can be clearly predicted. The industry says so,” he said.

In total, 10 units have been approved under the ISM. On manufacturing timelines for these plants, the minister said four semiconductor units, namely Kaynes, Micron, Tata, and CG Semi, are expected to begin commercial production in 2026.

He added that the Assam unit would start pilot production in the middle of the year, with commercial production by the end of the year.

“Practically all major telecom, power, EV, automotive - all are going to source chips from us,” Vaishnaw said.