Revenue Threshold For Firms Bidding For Government's AI Infra Tender Halved

The revenue threshold for companies looking to apply for the government's AI infrastructure services tender has been halved, opening up the bidding procedure to smaller players.

In a corrigendum issued on Friday available on the GeM portal, the Ministry of Electronics and IT said the primary bidder can have an average annual turnover of more than Rs 50 crore for last three financial years, down from the earlier requirement of Rs 100 crore.

However, preference will be given to those with an average annual turnover. As part of the competitive tendering process, for every Rs 5 crore over Rs 50 crore threshold, a bidder will get an additional 1 mark, upto a maximum of 10 marks. Marks are a part of the tender evaluation process to ensure that the best value for money is selected. 

Additionally, instead of a plain undertaking requirement earlier, bidders should now show a suitable bank guarantee and a purchase order for the number of anticipated GPUs or graphics processing units they have already paid for or are in the process of paying for. This could also point to the fact that the government is looking at players showing more intent for the tender.

As NDTV Profit had earlier reported, the government has inserted a new 'Make In India' clause, adding that the components used in providing AI cloud services by bidders should be procured from either a Class I local supplier or a Class II local supplier, in accordance with the 'Make in India' initiative guidelines issued by the Department for Promotion of Industry and Internal Trade.

MeitY had pushed out an RFP in August, inviting expressions for interest and empanelment from companies.

Companies that are in the business of data centres and cloud service providers are now expected to submit final bids to provide high-speed computing AI infrastructure like GPUs, accelerators, tensor processing units, storage to academia, startups, researchers, government bodies and others at the lowest rate that will be discovered through the bidding process.

The GPUs are a unit of compute, which is an important and expensive element in building AI systems, making it difficult to procure for smaller companies.

The AI compute infrastructure is one of the seven pillars of the government's IndiaAI mission, which was approved in March. It focuses on creating a scalable AI computing ecosystem with over 10,000 GPUs through public-private partnerships to meet the demands of India's AI startups and research community.