The government’s AI programme aimed at creating computing capacity under public-private-partnership mode through GPU (graphics processing unit)-based servers, has moved to the final stage.
Sources told FE that three firms, NxtGen Datacenter, E2E Networks, and Jio Platforms have emerged as the lowest bidders for supplying GPUs under the Rs 10,000-crore IndiaAI mission.
While NxtGen is understood to be the lowest bidders for supplying various GPU models quoted in the tender, for some GPU models Jio and E2E have emerged as the lowest bidders.
The announcement of the names is expected to be made on Thursday.
Once the tenders are awarded and GPUs procured and supplied to startups etc, an AI dataset platform will be created which will have the largest collection of anonymised data which will drive innovation and enhance capabilities of AI applications.
The GPU tender saw close contest between the three operators as the prices quoted by them for different GPU models were somewhat similar, sources said, adding that the other bidders will also get the opportunity to match the prices quoted by the lowest bidder and get empanelled.
A total of 19 bidders had participated in the bids, of which 13 firms qualified for the technical round where they showcased their readiness in several key areas, including GPU capacity, self-service portal to manage applications of beneficiaries, and their ability to provide the minimum quantity of GPUs at the earliest. Finally, the race narrowed to 10 bidders in the commercial round.
The government has earmarked Rs 5,000 crore of the total Rs 10,000 crore AI mission for creating computing infrastructure. A viability gap funding mechanism will be put in place to subsidise computing costs by up to 50% for startups, researchers, and academic institutions.
Sources said that NxtGen quoted one-fourth of the market price to make available the computing and most of the bidders priced it 10 times higher than the lowest. Jio and E2E quoted prices similar to the lowest bidder and are expected to match its rate.
Preference for computing infrastructure utilisation will be given to the lowest bidder, followed by the next lowest, ensuring fair and cost-effective access to resources. The empaneled companies will provide GPUs and related infrastructure, which will be available on an hourly basis to beneficiaries like startups, researchers, and academicians.
A major highlight of the IndiaAI mission is the creation of a robust AI compute infrastructure comprising 10,000 GPUs. This facility is crucial for startups and researchers to test and deploy large language models (LLMs) and other AI solutions.
As part of the empanelment requirements, companies must install a minimum of 1,000 AI compute units within six months of signing the agreement.
