The Centre has picked Bengaluru-based GenAI startup Sarvam AI to build India's first homegrown sovereign large language model (LLM) under the IndiaAI Mission.
Sarvam said in a statement that it will receive a dedicated compute infrastructure to build the model from scratch. The model will focus on reasoning, support voice-based tasks and work fluently across Indian languages. It is being designed for secure, population-scale use.
Notably, the LLM will be fully developed, deployed and optimised within the country, using local talent and infrastructure.
"…Our goal is to build multi-modal, multi-scale foundation models from scratch. When we do, a universe of applications unfolds. For citizens, this means interacting with AI that feels familiar, not foreign. For enterprises, this means unlocking intelligence without sending their data beyond borders," said Sarvam cofounder Vivek Raghavan.
Founded in 2023 by Raghavan and Pratyush Kumar, Sarvam AI is a full-stack GenAI platform that develops small and large language models.
It counts UIDAI, Urban Company, NITI Aayog, and the Ministry of Skill Development among its users.
In December 2023, the startup raised $41 Mn in its Series A funding round led by Lightspeed Venture Partners, along with participation from Peak XV Partners and Khosla Ventures.
Notably, The IndiaAI Mission, approved by the Union Cabinet last year, is eyeing to boost India's AI ecosystem by investing in compute capacity, research talent, datasets, applications, skills, startups, and trusted AI practices.
A key part is the IndiaAI Compute Capacity initiative, which will deploy over 10,000 GPUs to make AI infrastructure affordable and accessible to startups, researchers, and institutions.
As part of this push, the government has recently assigned GPU supply contracts to Yotta Data Services, E2E Networks, and NxtGen Cloud Technologies.
These companies will offer GPU access at subsidised rates, less than $1 per hour, said to be the lowest globally. Earlier reports suggested that Sarvam is also set to receive GPU resources worth INR 200 Cr to support its model development.
The Ministry of Electronics and IT was also said to be working on a large-scale facility with more than 18,700 GPUs, almost two-thirds of ChatGPT's capacity, to support domestic AI efforts.
Furthermore, the government also eased the eligibility criteria for the procurement of GPUs under the IndiaAI mission to allow more participation of smaller companies and startups.