New Delhi, September 10: Yotta Data Services plans to deploy around 40,000 Nvidia Vera Rubin GPUs at a new 120 MW data centre being developed at its Greater Noida campus, as the Indian data centre operator prepares for more than $12 billion of investment in artificial intelligence computing infrastructure over the next 12 months.
The Vera Rubin deployment will be housed at Yotta’s upcoming D4 data centre in Greater Noida, co-founder, managing director and CEO Sunil Gupta told Moneycontrol. Construction of the 120 MW facility is set to begin as the company expands its AI infrastructure footprint in India.
Yotta is separately planning to install around 40,000 Nvidia GB300 GPUs at its upcoming NM2 data centre in Navi Mumbai, an 80 MW facility.
Together, the Greater Noida and Navi Mumbai deployments would involve about 80,000 GPUs, with Gupta estimating the required investment at more than $12 billion over the next 12 months.
The planned spending would be separate from the capital Yotta intends to raise through its proposed initial public offering.
Gupta said the company plans to finance the GPU purchases through mechanisms including revenue-sharing agreements, GPU leasing and special-purpose investment structures with global investors rather than relying primarily on IPO proceeds.
Yotta is targeting a public listing in the January-March quarter of 2027. Moneycontrol reported that the company is considering an IPO of around $1.5 billion to $2 billion, while Reuters earlier reported plans to raise up to $1.5 billion through the offering.
According to Gupta, IPO proceeds would primarily be used to develop large data centres and sovereign cloud infrastructure rather than fund the entire GPU acquisition programme.
The company has already used alternative financing structures for its AI infrastructure expansion. Gupta said Yotta’s previously planned acquisition of 20,000 GPUs was backed through a roughly $2 billion arrangement with Nasdaq-listed Gorilla Technology involving revenue sharing from customers.
The latest plan represents a substantial expansion from Yotta’s previously announced GPU deployments.
In February, the company said it would deploy 20,736 liquid-cooled Nvidia Blackwell Ultra GPUs across its infrastructure, including its 60 MW D2 data centre at Greater Noida, with an investment exceeding $2 billion. Yotta said at the time that the deployment would form one of Asia’s largest Nvidia Blackwell HGX B300 superclusters.
The company’s earlier disclosures also outlined plans for tens of thousands of Nvidia GPUs across its Greater Noida and Navi Mumbai campuses as it builds infrastructure for increasingly compute-intensive artificial intelligence workloads.
Yotta’s Greater Noida presence dates back to 2022, when it opened its D1 hyperscale data centre and signed an agreement with the Uttar Pradesh government outlining investment of ₹39,000 crore over five to seven years for its data centre campus, IT equipment and associated infrastructure.
The proposed D4 facility would add another 120 MW of capacity to the campus and is intended to house Nvidia’s next-generation Vera Rubin GPUs.
The expansion comes as Yotta prepares to tap public markets to finance its broader growth. The company raised $150 million in primary capital in July at a valuation of about $3.9 billion and is in the final stages of raising another $150 million to $200 million ahead of its IPO, according to Moneycontrol.
Yotta is also in discussions with large private equity investors for additional funding and expects to file its draft red herring prospectus as early as October, subject to its fundraising and listing plans remaining on schedule.
The planned 80,000-GPU deployment across Greater Noida and Navi Mumbai would significantly increase Yotta’s AI computing capacity as demand for high-performance infrastructure accelerates among enterprises, cloud companies and developers building large artificial intelligence models in India.
Read also: Tata Steel deploys India’s first coke oven gas injection system at Meramandali as it targets sustainable steelmaking
