Govt spending on AI makes India lucrative for cloud providers

This is because these cloud service providers account for the bulk of the 34,333 GPUs procured by New Delhi, which in turn has given access to startups such as Sarvam, Gnani.ai and Soket AI Labs. GPUs are the fastest and most efficient way for companies to run calculations, allowing cutting-edge AI firms to analyse enormous amounts of data and train algorithms that power AI applications.

Yotta, for instance, is currently supplying nearly 17,000 of 34,333 GPUs to New Delhi.

“The India AI Mission is a key initiative that is helping us ramp up revenue as we look for funding avenues to complete our GPU orders and enable their access through our cloud platforms. Overall, we expect nearly 70% of our revenue to come from the Indian government’s procurement of GPUs from us,” said Sunil Gupta, chief executive of Yotta.

On 30 January, the ministry of electronics and information technology (Meity) announced the first tranche of GPU procurement, enlisting Jio Platforms, Tata Communications and Yotta Data Services, among others, in a total of 10 vendors supplying 18,693 GPUs to the AI Mission. On 30 May, Meity’s second tranche of GPU procurement added 15,640 more GPUs to what Ashwini Vaishnaw, Union minister for IT, said is “a one-of-a-kind central compute repository.”

Indian providers score

Yotta projected its fiscal year 2025 (FY25) revenue to be at $143.3 million,Mint reported on 10 February. Speaking withMint, Gupta said the company expects a major fillip driven by government-backed uptake of AI cloud infrastructure. “Despite us offering subsidized GPU pricing, India’s uptake of GPUs has been limited. In the long run, we expect to see a 5x boost to operating revenue by FY28, and a large part of it will come through government spending in India,” he said.

Tata Communications, too, is seeing a similar boost to its India operations. Neelakantan Venkataraman, vice-president and global head, edge and cloud business, Tata Communications, said the government-driven push is “definitely driving heavy demand, and we’re adding more GPUs as we speak to support the mission and cash in on it.”

Even beyond the India AI Mission, there is increasing enterprise AI adoption even in India. “Many pilots that began 12-18 months ago are moving into production stages, even if their scale hasn’t fully taken off yet. As a result, in India, there is a significant latent demand for AI infrastructure to fulfil the demand from enterprise data pipelines and support the creation of foundational models that the government has extensively spoken about,” Venkataraman sai

Data sourced from the company said that in FY25, Tata Communications earned $2.7 billion in gross revenue—growing at around 10% year-on-year. Revenue from cloud services contributed nearly 8% to the company’s annual revenue, but grew at a faster clip—at nearly 13% in FY25. Venkataraman also affirmed that Tata Communications already draws 42% of its overall revenue from India, and is expected to see a faster pace of growth in the coming fiscals—driven by the demand for AI infrastructure, especially from the government.

Jio Platforms, which is also supplying 1,000 GPUs to the Union government, had yet to respond toMintuntil press time.

Tendering process issues

Industry analysts, too, believe that government spending on AI infrastructure marks a big boost for India’s nascent AI industry. Sanchit Vir Gogia, chief executive at consultancy firm Greyhound Research, cited internal market research data to state that “68% of digital infrastructure executives in India now cite the India AI Mission as their first opportunity to win long-term, production-grade AI cloud deals with predictable utilization.”

This, though, is not without risk. Jayanth Kolla, co-founder and partner at fellow consultancy firm Convergence Catalyst, said a big risk to the ramp-up of government spending in AI “is the red-taped, bureaucratic government procuring process, and the restriction that comes with tenders.”

“The major vendors selected as infrastructure suppliers in the India AI Mission have large conglomerates backing them—which makes it easier for them to go through the tendering process. But, the very process of this does not take into account the idea of nifty innovation. While the rise in government spending is definitely great for India in the long run, an overhaul of the process is a necessity for India to reap the full benefits of the global AI rush,” Kolla added.

In the long run, though, there is strong scope for growth. Chirajeet Sengupta, managing partner at technology research firm Everest Group, said that despite the challenges, a worldwide push for sovereign AI initiatives will be the biggest boost for India’s on-cloud AI infrastructure providers.

“The Pentagon signed a deal with OpenAI with a token initial amount of $2 million for a central AI infrastructure. While it is set to scale up to $200 million for now, in the long run, subject to successful execution, this deal can go up to $2 billion. For India and the US, China’s leapfrog with DeepSeek has been a big push. Government spending is undoubtedly a big fillip as the future of public services and defence infrastructure, as well as cyber sovereignty, is critically dependent on AI. If the government speeds up, revenue fillip is definitely going to come through in the near term,” Sengupta added.