Yotta & Nvidia Announce $3B Deal to Deploy 20,000+ Blackwell Ultra GPUs

Yotta Data Services on Wednesday announced plans to deploy 20,736 liquid-cooled Nvidia Blackwell Ultra GPUs, creating one of Asia’s largest AI superclusters.

The project involves an investment of more than $2 billion and is expected to become operational by August 2026. This deployment will place India among a small group of countries capable of hosting advanced, large-scale AI infrastructure.

Highlighting closer India–US technology cooperation, Nvidia will establish one of the largest Nvidia DGX Cloud clusters in the Asia-Pacific region within Yotta’s HGX B300 Blackwell Ultra supercluster.

This deployment, part of a four-year agreement worth over $1 billion, expands an existing relationship, as Nvidia’s DGX Cloud has already been using Yotta’s GPU infrastructure over the past year. The expansion reflects growing demand for AI computing capacity both regionally and globally.

On the partnership, Jensen Huang, CEO of Nvidia said, “India is emerging as one of the world’s most important AI markets, driven by extraordinary talent and a bold national vision. Yotta’s deployment of one of the largest Nvidia Blackwell Ultra superclusters creates advanced AI infrastructure capable of training frontier-scale models and delivering AI at population scale. Expanding AI Factory capacity in India strengthens Nvidia’s regional footprint while supporting India’s ambition to build secure, sovereign, and globally competitive AI.”

AI Infra Trend

The partnership also reflects a broader global trend in which advanced AI infrastructure is being distributed across trusted countries. India’s emergence as a major AI infrastructure hub strengthens its strategic technology partnership with the United States and supports shared goals of building secure, high-performance AI ecosystems.

Yotta’s Blackwell supercluster will be built using Nvidia’s reference architecture and will include high-speed 800 Gbps Nvidia Quantum-X800 InfiniBand networking, advanced liquid-cooling systems, and more than 40 petabytes of high-performance storage. The system is designed to support the training of extremely large AI models with trillions of parameters, as well as high-volume inference workloads capable of handling millions of simultaneous prompts.

Yotta said its AI infrastructure, known as AI Factories, is designed to deliver faster deployment, lower operating costs, and strong performance for AI workloads. This infrastructure will allow global AI developers, enterprises, and governments to run advanced model training and critical AI applications reliably.

Yotta’s Shakti Studio

The company is also enhancing its Shakti Studio AI platform by integrating Nvidia Nemotron open models, Nvidia NIM microservices, and the full Nvidia AI Enterprise software suite. Through Shakti Studio, developers in India will gain access to open AI models, including their weights, training data, and development tools.

This will allow them to customise models, fine-tune them, and build sovereign AI solutions. The platform will help startups, businesses, and government organisations develop secure, production-grade AI applications using world-class infrastructure.

In addition, Yotta will allocate more than 10,000 Nvidia B300 GPUs from the supercluster to the IndiaAI Mission. This will support the development of Indian foundation models, research programmes, startups, and large-scale public AI platforms. This ensures that India’s domestic AI priorities grow alongside global AI infrastructure expansion.

The development supports India’s broader goal of building AI “from India, for India, and for the world.” By providing access to large-scale AI infrastructure within the country, it reduces reliance on overseas computing resources and enables Indian companies to build and deploy AI products for both domestic and global markets.

Installation at New Facility

The supercluster will be installed at Yotta’s 60 MW D2 hyperscale data centre in Greater Noida, which can be expanded to 250 MW.

It will also be supported by the company’s Navi Mumbai data centre campus, which can scale up to 2 GW. With dedicated power infrastructure, green energy sourcing, and integrated engineering capabilities, Yotta aims to build a long-term platform that could scale to more than one million GPUs over the next three to five years as India’s AI ecosystem grows.

Overall, the combined investments, including over $2 billion in infrastructure and more than $1 billion in contracted DGX Cloud capacity, highlight strong and sustained demand for advanced AI computing infrastructure in the region.