Why Are AI Giants Betting On India?

Last month, when Anthropic CEO Dario Amodei visited India, all eyes were on him. As he hopped around the country, meeting public officials and enterprise partners, one meeting that caught everyone’s attention was his sit-down with Indian Prime Minister Narendra Modi. Emphasizing India’s importance, Amodei posted after his meeting with Modi that how India deploys AI across sectors will be essential in shaping the future of AI. The Amazon-backed startup also announced that it will open an office in the southern Indian city of Bengaluru early next year.

Just a week after this development came another announcement: Google partnered with Adani Enterprises to invest $15 billion in building an AI data centre in the country.

These aren’t isolated incidents. Earlier in the year, OpenAI CEO Sam Altman met both Modi and India’s IT minister Ashwini Vaishnaw during his visit to India, later announcing plans for an office in New Delhi and a free ChatGPT Go subscription for Indian users for one year. At the same time, search-focused AI startup Perplexity partnered with India’s telecom operator Bharti Airtel to offer its subscription for free for 12 months to the 360 million Indians who subscribe to Airtel, while also adding more features for local users.

All of this comes at a time when India is gearing up to host the AI Impact Summit in February—a flagship global event where the Indian government wants to shape conversations around responsible, safe and inclusive AI. Many believe this could shift India’s role from a mere deployment ground to a part of the global AI ecosystem's structural architecture.

What’s driving the India push

India’s massive user base is the clearest driver of the current AI influx. Builders go where the users are, and India has quickly become the second-largest market for the major AI labs outside the US. ChatGPT usage in India has tripled since last year, and Claude sees a similar scale.

“India is, for the vast majority of these labs, among the top three to five markets by most metrics,” said Udbhav Tiwari, vice president of strategy and global affairs at Signal. “Scale matters.” Scale, he added, is not just about user numbers but linguistic and cultural breadth. India’s diversity makes it a “melting pot” for understanding what users want, he told Tech Policy Press.

But the draw goes beyond market size. “The chill in the United States on skilled immigration may be driving companies to open satellite offices closer to where the talent resides,” said Trisha Ray, associate director at the Atlantic Council’s GeoTech Center. “The Government of India (GoI) has been quick to seize the opportunity, signing MoUs on education, sovereign AI, and other priority areas. GoI's messaging has been strategic as well, folding partnerships into a broader narrative on AI-driven growth for all.”

US firms are also looking to derisk from regions exposed to trade tensions or regulatory uncertainty. “India has been one of those markets that’s been very conducive from an operational, regulatory and commercial perspective when it comes to AI,” said Amlan Mohanty, a lawyer and tech policy expert who is the lead writer of India’s AI governance guidelines, recently released by the government. “When it comes to governance and regulation, India’s been very pro-innovation — there’s no separate AI law like the EU has; and there’s been no strict enforcement actions on questions of privacy or data.”

In some ways, this mirrors earlier waves of Big Tech expansion. A decade ago, companies like Meta, Google, Amazon and Microsoft set up operations in India to tap its vast user base and engineering talent. “Many of the reasons remain the same: strong scientific engineering talent, good testing and fertile ground for high-impact applications (use cases) and closely integrated with other teams sitting out of India,” said Mohanty, who was at Google when it opened its India research lab in 2019.

But the similarities end there. Unlike previous waves in technology, AI is far more compute-intensive, and companies can’t rely solely on US-based infrastructure to serve India’s future demand. That has triggered major investments in domestic data centers. Google and Adani are putting $15 billion into an AI data center; OpenAI plans India data centers as part of its global $500 billion Stargate project; Amazon Web Services has committed $8.4 billion and begun construction in the state of Maharashtra.

“The interesting thing about AI companies coming to India is the broader infrastructure play,” said Divij Joshi, a research fellow in digital societies at ODI Global. “These investments are really about massive data centers.”

Tiwari said this trend is evident in the job postings and early hires AI labs are making in India. Unlike Big Tech’s earlier expansion focused on engineering and research, AI labs are prioritizing partnerships, policy and integration roles, he said. OpenAI’s first India hire was Pragya Misra on public policy and partnerships. Anthropic and Perplexity have posted similar roles. “The kind of functions that the AI labs are setting up in India right now aren't really hardcore development functions where you're working on core technology. Open AI’s first office is going to be in Delhi, not even in Bengaluru, because it is a partnerships and government-relations play,” said Tiwari. “This is very different from when Big Tech set up shop a decade ago.”

India wants to build Sovereign AI

For over a year now, India has been trying to position itself as a “global leader in artificial intelligence.” In March 2024, it launched the $1.2 billion IndiaAI Mission, aimed at expanding domestic compute capacity, building datasets, and supporting startups developing foundation models. One of the major parts of this program, the government had said, was to set up 10,000 GPUs in a bid to build compute capacity for the startups they’d be funding. One state minister described the mission as a way to “catalyze India’s AI ecosystem and position it as a force shaping the future of AI for India and for the world.”

At the heart of this sovereignty push are startups like Sarvam AI, which is building an India-focused large language model designed around local culture, diversity and languages. Sarvam already offers one of the largest open-source Hindi language models. An Indian minister recently said the country’s sovereign AI model will be ready by the time of the February summit.

India is not alone. Vietnam’s AI plans revolve around “sovereignty of national data and infrastructure.” Singapore, Malaysia, the UK and several European countries are also developing sovereign AI strategies.

India’s approach has some distinctive elements, according to experts involved in the AI summit preparations. These differences span development, adoption and regulation. “On adoption, the focus has been on developing socially beneficial use cases in public service delivery, education, agriculture, local language tools, and healthcare,” said Shruti Mittal, a research analyst with Carnegie Endowment for International Peace India, a think-tank, which has been organizing pre-summit events. “On capacity building and institutions, India is focusing on democratizing access to computing and decentralizing innovation by setting up AI labs in tier 2 and 3 cities with an emphasis on local language models.” Regulation, Mittal said, has shifted from an early instinct to regulate aggressively to a more measured position that avoids stifling innovation while prioritizing safety and risk mitigation. This is reflected in the Ministry of Electronics and Information Technology’s new AI Governance Guidelines.

Sovereignty runs through all these choices. “New Delhi's stance on AI is animated by sovereignty at its core, and can be characterized as a hybrid approach, with the government partnering with industry to co-develop AI infrastructure and use cases,” said the Atlantic Council’s Ray. “The government steps in selectively — focusing on key sectors or public needs — while letting markets do the rest.”

This has produced a blend of homegrown open-source tools like Bhashini, large open datasets, and national compute infrastructure — boosted by partnerships with Microsoft, IBM and Nvidia, she said.

For India, many say, Sovereign AI seems to be about “self-reliance” and “strategic autonomy” to ensure that the AI systems developed or deployed reflect the country’s priorities and diversity. “Sovereign AI does not mean isolation from global players; it means ensuring that India retains control over how AI is built, used, and scaled within its borders,” said Rameesh Kailasam, CEO of startup and tech policy advocacy group IndiaTech. He further believes that with AI becoming central to geopolitical competition, India has positioned itself as a “trusted, democratic, and neutral destination amid competing global regulatory models.”

What happens if the bubble bursts?

The AI Impact Summit and India’s bet on shaping global conversations around AI come at a time when there are concerns of a potential AI bubble burst. “The whole thing seems to be propped up by a lot of speculation,” said Joshi. “Policy seems to be following hype and capital, rather than directing AI toward clear social goals.”

Others argue India is insulated. “The bubble, if it bursts, is largely about unsustainable infrastructure investments,” said Mohanty. “India is prioritizing applications, small, efficient models that don’t depend on massive infrastructure.”

Tiwari said a bubble before February is unlikely, but early signs could shape the summit. “There could be questions around what role the government should play, or banks and financial institutions play, in order to prop up strategic national assets,” said Tiwari.

For now, AI giants continue to bet big on India, while the government prepares to host the summit in the coming months, positioning itself as a global AI leader.