The last phase of the government’s AI program focuses on utilizing GPU-based servers to build computing capacity through a public-private partnership.

The lowest bidders for GPUs under the Rs 10,000-crore IndiaAI mission are NxtGen Datacenter, E2E Networks, and Jio Platforms.
Government’s AI-Based Program To Focus on GPU-Based Servers: NxtGen, E2E, Jio To Supply GPUs
Jio and E2E are the lowest bidders for some GPU models, while NxtGen is the lowest bidder for a variety of GPU models.
On Thursday, the lowest bidders are anticipated to be formally announced.
An AI dataset platform will be developed following the awarding of the tenders and the purchase of GPUs, with the goal of gathering anonymized data to spur innovation and improve AI capabilities.
Because the three companies’ quotes for the GPU tender were comparable, there was a fierce competition. There will be an opportunity for other bidders to match the lowest prices.
Thirteen of the 19 bidders who entered the tender made it to the technical round, where they proved themselves prepared in crucial areas like GPU capacity, self-service portals, and speedy GPU delivery.
Out of the Rs 10,000 crore budget for the AI mission, the government has set aside Rs 5,000 crore for the development of computing infrastructure.
Startups, researchers, and academic institutions will have access to up to 50% funding for computing costs through the introduction of a viability gap funding mechanism.
NxtGen’s Quotes Only A Quarter of Market Price
While the majority of other bidders provided prices ten times higher than NxtGen’s, NxtGen’s quotes were only a quarter of the market price.
In order to stay competitive, Jio and E2E are anticipated to match NxtGen’s pricing.
In order to ensure cost-effective access to resources, the lowest bidder will be given preference for using the computing infrastructure, followed by the next lowest bidder.
For beneficiaries such as researchers, academicians, and startups, the affiliated companies will supply GPUs and associated infrastructure on an hourly basis.
Building a strong AI computing infrastructure with 10,000 GPUs is a crucial component of the IndiaAI mission since it is necessary for testing and implementing large language models (LLMs) and other AI solutions.
Companies must install at least 1,000 AI compute units in accordance with the empanelment requirements within six months of the agreement’s signing.