Out of the 506 proposals received by the India AI Mission for building foundation AI models, 43 are specifically dedicated to building LLMs, underscoring the nation's strategic emphasis on sovereign and culturally relevant AI, said Sunil Gupta, CEO of Yotta Data Services.

He was speaking at a panel discussion at the AI for India Summit 2025, organised by AI4India in Bengaluru.

The significant proportion of LLM proposals, each requesting substantial GPU resources (more than 2,000 GPUs per model), highlights a pivotal shift in India's AI landscape, Gupta said.

Just 18 months ago, the debate was centred on whether India needed even 1,000 GPUs, he said.

“At least 47 foundation model proposals apparently need more than 2,000 GPUs, which makes us optimistic for having offered Nvidia Blackwell GPUs in the second round of India AI Mission's GPU tender,” Gupta told ET on the sidelines of the event.

As part of the mission, the government is incentivising the development of LLMs built by startups such as Sarvam, Gnani, Gan and Soket AI Labs with investment capital and other support.

According to people familiar with the matter, the IndiaAI Mission will soon announce more LLM-developing applicants who will be given GPU access or grants. Likely contenders include conversational AI platform startup CoRover.ai, a consortium of Indian professors based abroad and an agricultural research university.

ET recently reported that out of the proposed capacity of 34,333 GPUs in the IndiaAI Mission, the installed capacity is 17,374 GPUs.

The current demand reflects a robust commitment to creating indigenous foundational models, driven by geopolitical considerations of data sovereignty and the need for AI to adapt to India's diverse cultural and linguistic contexts.

"With Bhashani, I've seen their scale increase from about two million inferencing per day to about 16 million per day," Gupta said at the panel discussion.

While a significant portion of the funding for model creation currently stems from government grants or private equity, the focus remains on fostering large-scale development that can cater to India's unique needs.

This push is expected to pave the way for a new era of AI applications, especially as use cases for inferencing at a population scale emerge across various sectors.