IndiaAI Mission enhances AI infrastructure with 14,000 GPUs for compute capacity

Over 10,000 of the 14,000 GPUs secured for the IndiaAI Mission are available through providers like Yotta Data Services, E2E Networks, Tata Communications, and AWS, while 4,000 more will be procured by Jio Platforms and CtrlS Datacenters

The IndiaAI Mission is taking significant steps toward enhancing the nation's artificial intelligence (AI) infrastructure by securing approximately 14,000 graphics processing units (GPUs) through competitive bidding. The move aims to provide crucial compute power for AI research and development, benefiting academia, startups, and government agencies. The mission also plans to roll out a subsidy scheme that will cover up to 40% of compute costs for eligible users, further accelerating the growth of AI in India.

Of the 14,000 GPUs secured, over 10,000 are already available through major providers like Yotta Data Services, E2E Networks, Tata Communications, and AWS’s managed service providers. The remaining 4,000 GPUs will be procured by companies like Jio Platforms and CtrlS Datacenters. While the initial proposals from the 10 bidding companies included a total of 18,693 GPUs, the final number was adjusted after companies matched the lowest bid rates, as outlined by the IT ministry on January 30, with final bids due by February 14.

High-end GPUs drive IndiaAI Mission

A majority of the GPUs, about 70%, will be high-end models such as Nvidia’s H100, while the remaining 30% will consist of older or lower-end GPUs. Yotta Data Services will provide the largest share, contributing 9,216 GPUs, including 8,192 Nvidia H100s. AWS's managed service providers will offer 1,200 low-end GPUs, while Jio Platforms will provide 208 Nvidia H200 GPUs and 104 AMD MI300X GPUs. The IndiaAI Compute Portal, soon to be launched by IT Minister Ashwini Vaishnaw, will streamline access to these GPUs for various stakeholders, including union ministries and state governments, facilitating AI resource allocation.

In addition to GPUs, the mission has secured competitive bids for storage and network services. NxtGen Datacenter will supply high-speed storage, while Yotta Data Services will offer object storage, and AWS MSPs and Jio Platforms will provide free data ingress services. E2E Networks will offer free egress network services.

With a budget of Rs 10,371.92 crore (USD 1.25 billion), the IndiaAI Mission has allocated Rs 4,563.36 crore for compute capacity expansion over the next five years, marking a significant investment in India’s AI ecosystem.