AI computing infrastructure providers such as Locuz Enterprise Solutions, NxtGen Datacenter, E2E Networks, Jio Platforms, and CtrlS, will have a headstart in supplying graphics process units (GPUs) for the Rs 10,000 crore IndiaAI mission, as they have emerged as the lowest price bidders (L1) for supplying the same.
Only once their capacity is exhausted, the other empanelled players will get a chance to supply GPUs. The government has empanelled a total of 10 companies, which also includes Yotta, Tata Communications, CMS Computers India, Orient Technologies, and Vensysco, to provide 18,693 GPU units.
Barring Yotta, which has proposed 9,216 GPUs, the other bidders have proposed to make available nearly 1,000 GPU units. For companies like Yotta who are lowest second (L2) or third bidders (L3), the opportunity for their GPUs consumption, however, would come in after the other lower bidders exhaust their capacities or are unable to make available their GPU units.
According to the details shared by the government, out of 18,693 GPUs offered by the bidders across 30 models, 12,896 are Nvidia H100 GPUs, 1,480 are H200, 512 units are AMD Mi325X, and 230 units are Mi300X.
Locuz emerged as the lowest bidder in 10 GPU models, followed by NxtGen in seven, E2E and Jio in six models each, and CtrlS in one model.
Average rate per computing unit is Rs 115.85 per hour, whereas average rate per higher precision AI computing unit is Rs 150 per hour. The government said the average discount from market prices for AI Computing services is 42%. Higher precision AI computing units, critical for foundational model training, are being provided at a discount of 47% from market prices.
FE was the first to report that NxtGen, Jio, and E2E are the lowest bidders for different GPU models.
“The IndiaAI empanelment provides a profitable business opportunity for us. Our supplies for GPUs are expected over the next few days. Over the next four years, this could be a Rs 5,400 crore revenue opportunity for us,” said AS Rajgopal, managing director and chief executive officer of NxtGen Datacenter and Cloud Technologies.
The government said the empanellment process will be continuous, and other entities can also be part of the process. Infact, it will soon issue a notice to procure more GPUs.
Notably, the government has earmarked Rs 4,563 crore out of Rs 10,000 crore for creating the AI computing infrastructure. It will implement a viability gap funding model, subsidising computing costs by 40% for startups, researchers, and academic institutions.
