A portion of those GPUs will be used by the government as part of its India AI Mission, while the rest may be used by global model builders and hyperscalers who are keen to double down their presence in the country.
“I have committed to providing 17,000 H300 GPUs to the government for the next leg of the AI Mission. There is demand because artificial intelligence (AI) is being adopted for population-level scaling of infrastructure. There’s demand from global players as well,” he told Business Standard on the sidelines of the India AI Impact Summit.
The investment is on top of the $2 billion that was announced by Yotta on Wednesday to deploy 20,736 GPUs at its data centre in Noida. Nvidia will also establish one of Asia Pacific’s largest NVIDIA DGX Cloud cluster within Yotta’s supercluster, under a four-year engagement valued at over $1 billion.
Gupta said he is yet to firm up plans on how he wants to fund the mammoth investment, but options include a mix of equity and debt. He is looking to raise about $1 billion through a mix of pre-initial public offering (IPO) and IPO while for the rest $3 billion, going the debt route may be an option.
The company has already announced plans to list in India in 2026-27 (FY27) — after steering away from a potential US listing — as demand in the country soars.
The demand comes at a time when the government has decided to expand its compute capacity under the AI Mission by another 20,000, besides the 38,000 already there as part of the first leg. While the first leg had an investment of ₹10,300 crore, the second part of the Mission is expected to double the outlay.
Yotta, owned by the Hiranandani family, builds cloud and AI infrastructure, and has several data centres in Mumbai, GIFT City in Gujarat, and Noida.
 "Sunil Gupta, MD and CEO, Yotta Data Services")