IndiaAI Mission

Imagine wanting to train an artificial intelligence model but discovering that renting the specialised computing power needed, banks of Nvidia GPUs, costs more per hour than most Indian startups can realistically afford, a cost barrier steep enough to keep AI development concentrated among a handful of well-funded companies globally. The IndiaAI Mission exists specifically to remove that barrier for Indian developers and startups.

Approved in March 2024 with a Rs 10,371.92 crore, roughly $1.25 billion, outlay over five years, the mission's core mechanism is straightforward, the government aggregates GPU capacity and makes it available to Indian startups, researchers and developers at heavily subsidised rates, roughly Rs 65 per hour, well below open-market pricing, specifically to democratise access to the compute power AI development genuinely requires.

The scale has grown rapidly, India's aggregated AI compute capacity crossed 38,000 GPUs by February 2026, with a further 20,000 added shortly after, and the government has committed to reaching 100,000 GPUs by the end of 2026, running across facilities including a live 30 MW cluster in Chennai and a 40 MW facility planned for Mumbai.

This mission connects directly to the sovereign AI models covered elsewhere on this site, BharatGen's Param2 and Sarvam AI's vernacular-language models were built substantially on compute made accessible through exactly this kind of subsidised access, meaning the IndiaAI Mission isn't just an infrastructure programme, it's the specific enabler letting Indian AI developers compete without needing the billion-dollar compute budgets that global AI labs command.