Nvidia and Yotta to Deploy $2 Billion AI Supercluster in India With 20,736 Blackwell Ultra GPUs Under IndiaAI Mission

India is about to host one of the world’s most powerful artificial intelligence computing facilities. Yotta Data Services, one of India’s leading data centre operators, has announced a partnership with Nvidia to deploy a $2 billion AI supercluster featuring 20,736 liquid-cooled Nvidia Blackwell Ultra GPUs — infrastructure that will rank among Asia’s largest AI computing installations when it goes live by August 2026. The announcement, made during the India AI Impact Summit in New Delhi, marks a decisive step in India’s ambition to build sovereign AI capabilities and reduce its dependence on overseas computing resources. For those following artificial intelligence developments in India, this deployment represents a quantum leap in available compute power.

Inside the $2 Billion Supercluster

The Yotta-Nvidia supercluster will be deployed at Yotta’s 60-megawatt D2 hyperscale data centre at its Greater Noida campus in Uttar Pradesh, with the facility scalable to 250 megawatts. Additional capacity will be supported by Yotta’s Navi Mumbai campus, which is scalable to a massive 2 gigawatts. The deployment involves Nvidia’s latest Blackwell Ultra architecture — the B300 series GPUs — which represent the cutting edge of AI training and inference hardware, offering significant performance improvements over the previous Hopper generation.

The 20,736 GPUs will form an HGX B300 Blackwell Ultra supercluster with liquid cooling, optimised for large language model training, scientific computing, drug discovery simulations, and inference workloads. Each B300 GPU offers approximately 2.5 times the training performance of the H100 for large language models, meaning the combined cluster will deliver computing power equivalent to several hundred thousand older-generation GPUs. This positions India to train frontier AI models domestically rather than relying on cloud computing resources hosted overseas.

A $1 Billion DGX Cloud Deal: Nvidia’s Largest in Asia-Pacific

Alongside the infrastructure deployment, Yotta has signed a four-year commercial engagement worth over $1 billion with Nvidia to establish one of the Asia-Pacific region’s largest Nvidia DGX Cloud clusters within the supercluster. Nvidia’s DGX Cloud service has been utilising Yotta’s GPU infrastructure over the past year, and the expanded deployment extends that relationship significantly.

The DGX Cloud model allows organisations — from startups to large enterprises — to access Nvidia’s most advanced AI computing resources on demand without purchasing hardware. For India’s rapidly growing AI ecosystem, this means that startups like those recently raising significant capital can access world-class computing power locally. The billion-dollar funding rounds flowing into Indian AI companies require corresponding compute infrastructure, and the Yotta-Nvidia partnership aims to meet this demand domestically.

10,000 GPUs Committed to IndiaAI Mission

Perhaps the most significant aspect of the announcement is the commitment of over 10,000 Nvidia B300 GPUs from the supercluster to the government’s IndiaAI Mission. This initiative, launched with a budget of Rs 10,372 crore, aims to build sovereign AI infrastructure that allows Indian researchers, startups, and government agencies to train and deploy AI models on Indian soil, ensuring data sovereignty and reducing latency for domestic applications.

The IndiaAI Mission has been building a network of AI computing resources across the country, but the Yotta contribution represents the single largest private-sector commitment to the programme. The GPUs will be available to approved researchers and organisations through the IndiaAI Compute Portal, dramatically expanding the amount of compute available for government-backed AI research in areas such as agriculture, healthcare, language translation, and national security.

Nvidia’s Broader India Strategy: L&T, Reliance, and Sovereign AI

The Yotta partnership is part of a much larger Nvidia engagement with India. During the India AI Impact Summit, Nvidia CEO Jensen Huang unveiled plans to deploy tens of thousands of Blackwell GPUs across the country through multiple partnerships. Engineering giant Larsen and Toubro is working with Nvidia to build a gigawatt-scale network of AI data centres, designed to keep critical data and model training within India’s borders — a key requirement for government and regulated industries.

Reliance Industries, through its Jio Platforms subsidiary, has also been expanding its AI infrastructure partnership with Nvidia. The broader picture is one of India rapidly building the physical infrastructure needed to participate in the global AI race at the frontier level. As Huang stated during the summit, artificial intelligence is driving the largest infrastructure buildout in human history, and every country will need to build sovereign AI capabilities. India, with its 1.4 billion population, deep engineering talent, and growing digital economy, is among the most important markets for Nvidia globally.

The deployment also has implications for India’s broader technology and science ecosystem. AI computing infrastructure supports not just commercial applications but also scientific research — from protein folding and drug discovery to climate modelling and materials science. With 20,736 Blackwell Ultra GPUs available in India, researchers who previously had to queue for overseas cloud computing resources or work with limited local hardware will have access to genuinely world-class compute.

The Road Ahead for AI Infrastructure in India

India’s total AI computing capacity is set to increase dramatically in 2026 and 2027. Between Yotta, Jio, Adani, and several other data centre operators, the country is on track to have over 100,000 advanced GPUs operational by 2027. Whether this is enough to match the compute needs of a country with India’s ambitions remains to be seen — China and the United States each have millions of high-end GPUs in operation. But the trajectory is clear, and the Yotta-Nvidia supercluster represents the most tangible proof yet that India is serious about becoming an AI power in its own right.