IndiaAI Mission: Large IT, AI firms may not get govt nod to make LLMs

Meity says these firms can obtain GPUs on their own

The government is unlikely to consider the applications submitted by large information technology service providers and well-funded artificial intelligence (AI) companies to build indigenous large language models (LLMs), sources told Business Standard. 

The Ministry of Electronics and Information Technology which is the nodal ministry overseeing the ₹10,372 crore IndiaAI Mission, believes that larger IT service providers or even AI companies can obtain the requisite graphics processing units (GPUs) or access funds on their own. 

“Taxpayers’ money cannot be given to these companies which can access the market and buy GPUs at a competitive price from any vendor. They are welcome to make the LLMs if they want to but it will not be fair to give them incentives under the (IndiaAI) mission,” a senior government official said. 

Last Friday, the IT ministry announced that three new indigenous large language models will be developed by Gnan,ai, Gan.ai, and Soket, in addition to the one already underway at Sarvam. 

Soket is set to build India’s first open-source 120-billion parameter foundation model, optimised for linguistic diversity and aimed at sectors such as defence, health care, and education. 

Gnani.ai will develop a 14-billion parameter voice AI model designed for real-time speech processing and advanced reasoning. Meanwhile, Gan.ai will work on a 70-billion parameter multilingual foundation model with a focus on “superhuman text-to-speech” capabilities, the ministry had then said. 

LLMs use vast data repositories and compute power to first learn and then be able to generate human-like responses to a user’s inputs, including generating a new idea. The GPU procurement effort and the LLM development initiatives are being carried out under the IndiaAI Mission. 

  The government is hopeful of clearing another “12-15 proposals from various other startups” to build indigenous LLMs, small-language models (SLMs), large multimodal models (LMMs), and large reasoning models under the IndiaAI Mission, another official said. 

Under the IndiaAI Mission, the government has so far received bids to buy and procure more than 34,000 GPUs, and is hopeful to get bids for another 12,0000 by June 10, when the bidding for the third round ends. 

In the first round, the government received bids from 10 companies to procure and supply 18,693 GPUs against the target of 10,000 GPUs. These bids were received from CMS Computers, CtrlS Datacentres, E2E Networks, Jio Platforms, Locuz Enterprise Solutions, NxtGen Datacentre and Cloud Technologies, Orient Technologies, Tata Communications, Vensysco Technologies, and Yotta Data Services. 

The second round of bidding to procure and supply GPUs under the IndiaAI Mission saw a total of 15,640 GPUs being added by seven bidders, which include Cyfuture India, Ishan Infotech, Locuz Enterprise Solutions, Netmagic IT Services, Sify Digital Services, Vensysco Technologies, and Yotta Data Services. 

Of the 18,693 GPUs that were bid for and procured in the first round, Sarvam has been allocated 4,000 GPUs for six months to work on the indigenous LLM it is developing, while another 3,000 GPUs have been allocated to the Bhashini division of the IT ministry for work on language translation, a senior government official said.