IndiaAI Mission: All firms match L1 prices, GPU access portal to launch in 7-8 days

Union IT Minister Ashwini Vaishnaw addressing the Press on February 15 in Bengaluru.

The government is set to launch the graphics processing unit (GPU) access portal within the next seven to eight days, minister for Electronics and Information Technology Ashwini Vaishnaw said in Bengaluru on February 15.

"Very soon, in the coming 7-8 days, we'll be launching the portal. So when we launch the portal, you'll come to know," said Vaishnaw.

This comes after L2 bidders selected for supplying the GPUs matched L1 prices, a senior government official told Moneycontrol. February 14 was the last date for matching the prices. Both L1 and L2 bidders were selected to supply the GPUs, the government had told on January 31.

L1 bidders include Jio Platforms, NxtGen Data Center, Locuz Enterprise, E2E Networks, CtrlS DataCenters.

With this, CMS Computers, Orient Technologies, Tata Communications, Vensysco and Yotta Data Services, all L2 bidders, are officially empanelled in the IndiaAI GPU supply program.

Also read: IndiaAI GPU procurement sees record discounts as firms slash prices by up to 89%

On February 17, Yotta, while announcing the same, said that it will provide 50 per cent of all the advanced GPUs under the mission, and that the same will be available on the upcoming GPU portal.

Of the 18,693 GPUs earmarked under the IndiaAI Mission, 12,896 are Nvidia H100 GPUs, 1,480 Nvidia H200 GPUs and rest are high-performance AI chips from AMD, Intel, and AWS.

These GPUs are going to be made available for startups and researchers on the portal, which Minister Vaishnaw said is going to be launched within seven to eight days.

Application programming interfaces (APIs) will have to be installed both at the suppliers' end and the subscribers' end to connect the two, the official said.

On January 29, Moneycontrol reported that the bidding took place on January 22, citing sources.

As much as 44 percent, or Rs 4,563.36 crore, of the Rs 10,371.92 crore of the IndiaAI Mission has been earmarked for providing compute capacity of more than 10,000 GPUs over five years.

This development comes at a time when Chinese startup DeepSeek with their open-source reasoning model R1 has challenged the narrative that entities need access to large amounts of GPU for building frontier models.

"IndiaAI's compute pillar is very open and flexible. Ultimately, the idea is that institutions and researchers use GPUs and bring in innovations," the official said indicating that DeepSeek's breakthrough in terms of efficiency will not have any detrimental impact on the government's AI infrastructure plans.

The official also noted that DeepSeek model was open-source and said that startups and researchers were free to leverage the model for building their innovations.

MeitY’s current approach of procuring GPUs in bulk avoids traditional viability gap funding, which could have excluded existing GPU capacity providers.

Also, read: Decoding IndiaAI: Inside India's Rs 10,372-crore plan to boost AI infrastructure

Update, February 17: Comments added from Yotta Data Services on being officially empanelled in IndiaAI Mission