Business
Yotta Bets Big On Nvidia’s Latest Chips To Build Asia’s Largest AI Supercluster
Last Updated:
Yotta Data Services to spend $2 billion to deploy Nvidia’s latest Blackwell chips in India, building one of Asia’s largest AI superclusters
Yotta Infrastructure’s Greater Noida data centre park (Photo Credit: Yotta’s website)
India’s data centre sector is entering a new era of scale. Yotta Data Services said Wednesday, February 18, that it will deploy Nvidia’s most-advanced artificial intelligence chips in a $2-billion project that will establish one of the largest AI computing hubs in Asia, positioning the country as a serious contender in the global race for AI infrastructure.
The investment centres on the first-ever deployment of Nvidia’s Blackwell B300 graphics processing units in India, to be housed at Yotta’s hyperscale campus in Noida, just outside New Delhi. The supercluster is expected to go live by August.
Anchoring the project is a four-year agreement with Nvidia valued at roughly $1 billion, under which the chipmaker will establish one of Asia-Pacific’s largest DGX Cloud clusters within Yotta’s infrastructure. Sunil Gupta, managing director and chief executive of Yotta, told The Economic Times that Nvidia will deploy approximately 10,300 GPUs through the arrangement to serve its global Asia-Pacific customers and run its own models and services. “Nvidia is creating one of Asia’s largest DGX Cloud clusters on our supercluster,” Gupta said.
The deal underscores a broader shift in how hyperscalers and chipmakers are approaching India. Global cloud providers, including Microsoft and Amazon, have been expanding AI data centre capacity in the country, drawn by surging demand for generative AI services and government pressure to localise advanced computing infrastructure, according to Reuters. Nvidia’s direct commitment within Yotta’s facility goes a step further, signalling confidence in India as a viable hub for serving enterprise AI workloads across the region.
A significant share of remaining capacity will be dedicated to India’s national AI Mission, which has received more than 500 applications from start-ups seeking affordable compute access. Gupta told The Economic Times that the expansion will increase the country’s compute capacity “by almost five to six times”, addressing what he described as enormous pressure on existing resources. The infrastructure will support state-backed Indian language model initiatives, including Bhashini, Sarvam, BharatGen and Soket, all aimed at building foundational AI models trained on Indian-language datasets.
Yotta currently holds around 10,000 advanced Nvidia GPUs, accounting for nearly 75% of India’s GPU compute capacity. With the new deployment, its total GPU count will rise from roughly 40,000 to more than 75,000 over the next two years.
The capital push is being funded through a combination of debt and equity. Speaking to CNBC-TV18, Gupta said the company is targeting a fundraise of close to $1 billion to support its current phase of GPU deployment. Yotta has already invested over $1.5 billion in infrastructure and expects to commit an additional $2 billion toward advanced chips. A pre-IPO equity round is underway, with the company aiming to enter public markets within the current financial year.
Yotta is part of Indian billionaire Niranjan Hiranandani’s real estate conglomerate and operates data centre campuses in Mumbai, Gujarat and near New Delhi. Additional capacity from its Mumbai facility will supplement the Noida supercluster.
The timing of the investment is notable. US export controls have reshaped global supply chains for advanced AI semiconductors, pushing technology firms to deepen partnerships in markets that remain accessible. India, which has cultivated strong ties with Washington and positioned itself as a neutral beneficiary of great-power competition in technology, has emerged as one of the cleaner plays for companies looking to expand AI compute outside China.
Speaking to CNBC-TV18, Gupta said that India’s AI ambitions are grounded in practical outcomes, with the goal of delivering impact across agriculture, healthcare, education and climate. He drew a comparison to the Unified Payments Interface, suggesting AI-led transformation could similarly reshape how services are delivered at scale across the country.
For Nvidia, the DGX Cloud anchor at Yotta is the latest in a string of sovereign and commercial AI infrastructure deals across Asia, as the company works to deepen its footprint ahead of any potential tightening of chip export restrictions.
Check JEE Mains Result 2026 Link Here
February 18, 2026, 11:30 IST
Read More