Yotta Data Services has announced a major expansion of its artificial intelligence (AI) infrastructure, outlining plans to scale its NVIDIA GPU capacity to nearly 85,000 GPUs, including the deployment of 20,736 liquid-cooled NVIDIA Blackwell Ultra GPUs. The announcement, made on 18 February 2026, forms part of the company’s strategy to strengthen its AI cloud platform for enterprise, hyperscale, sovereign AI and high-performance computing (HPC) workloads.
The expansion is backed by an investment of more than USD 2 billion and is expected to become operational by August 2026. The Blackwell Ultra supercluster will be deployed at Yotta’s 60 MW D2 hyperscale data centre in Greater Noida, which is scalable to 250 MW, while additional AI infrastructure capacity will be supported from the company’s Navi Mumbai campus, designed to scale up to 2 GW.
Yotta currently operates more than 10,000 NVIDIA GPUs in production and expects another 8,000 GPUs to go live within the next quarter, before adding the new Blackwell Ultra deployment. The company said it has a roadmap to scale beyond 80,000 GPUs by FY27, supported through phased infrastructure expansion and long-term capacity planning.
According to the company, the expanded AI infrastructure will support large language model (LLM) training, AI inference, generative AI applications and other compute-intensive enterprise workloads. The Blackwell Ultra supercluster will feature liquid cooling, 800 Gbps NVIDIA Quantum-X800 InfiniBand networking, and more than 40 PB of high-performance parallel file-system storage, enabling large-scale AI model development and inference.
As part of the initiative, NVIDIA will establish one of the Asia-Pacific region’s largest DGX Cloud clusters within Yotta’s Blackwell Ultra supercluster under a four-year engagement valued at more than USD 1 billion. The collaboration is intended to expand access to advanced AI infrastructure for enterprises, developers and research organisations while strengthening India’s AI computing ecosystem.
Yotta said a significant portion of its existing GPU infrastructure has already been contracted by customers in the United States and Europe, reflecting growing international demand for AI compute capacity as organisations continue to invest in foundation models, generative AI and enterprise AI applications.
Commenting on the announcement, Sunil Gupta, Co-Founder, Managing Director and CEO of Yotta Data Services, said, “India’s AI ambition requires sustained, high-performance compute at scale. By combining Blackwell Ultra infrastructure with open models like NVIDIA Nemotron and the full NVIDIA AI stack, we are enabling developers to build sovereign, globally competitive AI applications from India.”
Darshan Hiranandani, Co-Founder and Chairman of Yotta Data Services, said the company’s capital strategy is focused on building scalable AI infrastructure that supports both national priorities and international demand for AI compute.
To support its continued expansion, Yotta plans to raise USD 1.25 billion to USD 1.5 billion through a pre-IPO funding round followed by an initial public offering. The company has already secured USD 150 million from approximately 150 high-net-worth individuals and family offices as part of its funding strategy.
Yotta stated that its cumulative investment commitment towards AI infrastructure now stands at approximately USD 7 billion, covering GPU clusters, hyperscale data centres, cloud infrastructure and AI computing platforms.
The announcement comes amid growing investments in AI infrastructure globally as enterprises increase spending on AI model training, inference and next-generation digital workloads. Technology providers and cloud operators are expanding GPU-based compute infrastructure to meet rising demand for sovereign AI capabilities and enterprise-scale AI deployment.
Yotta Data Services is an Indian data centre and cloud infrastructure company providing hyperscale data centres, GPU-as-a-Service (GPUaaS), AI cloud platforms and enterprise cloud solutions. The company focuses on building large-scale AI infrastructure to support enterprises, startups, research institutions and public-sector organisations across India and international markets.
 infrastructure, outlining plans to scale its NVIDIA GPU capacity to nearly 85,000 GPUs, including the deployment of 20,736 liquid-cooled NVIDIA Blackwell Ultra GPUs. The announcement, made on 18 February 2026, forms part of the company's strategy to strengthen its AI cloud platform for enterprise, hyperscale, sovereign AI and high-performance computing (HPC) workloads.){kind=link}