The Indian government is examining a dedicated financing vehicle for frontier artificial intelligence companies and compute infrastructure, with a proposed anchor commitment of ₹15,000 crore to ₹20,000 crore. The National Frontier AI & Compute Fund, or NFAICF, is being considered under the IndiaAI Mission to provide long-duration risk capital for foundational-model developers, GPU clusters, specialised data centres and other capital-intensive parts of the AI stack. The proposal remains at the consultation stage: its final size, governance structure, investment model and level of government participation have not been decided.
A concept discussed with ecosystem stakeholders contemplates using government capital as an anchor to mobilise private investment at a ratio of three or four rupees for every rupee committed by the state. A regulated investment structure is among the options under consideration. The design recognises that frontier AI ventures require investment periods extending beyond conventional venture-capital cycles: a 12-year fund life with a possible two-year extension has been discussed, compared with the five-to-seven-year horizons common in technology investing.
The proposal was discussed at a closed-door meeting in New Delhi on September 24 involving IndiaAI Mission officials, startup founders, investors and infrastructure companies. Participants included Ministry of Electronics and Information Technology Secretary S Krishnan, Sarvam AI co-founder Vivek Raghavan and Peak XV Partners managing director Rajan Anandan. The discussions covered the financing requirements of companies building large models, domestic compute capacity and advanced AI infrastructure.
The financing requirement is being framed around the rapidly increasing cost of developing frontier systems. Material circulated during the consultation cited estimates that individual training runs for advanced models could exceed $1 billion by 2027. Such expenditure would be difficult to finance through grants or conventional venture rounds, particularly when companies must also secure GPUs, power, data-centre capacity, engineering talent and repeated model-training cycles.
The fund would complement the existing IndiaAI Mission architecture, which provides subsidised access to shared compute and supports datasets, applications, startups and foundational-model development. The government has previously said that more than 38,000 GPUs had been made available through the common-compute programme, with another 20,000 planned. The proposed NFAICF would address a different financing layer by supplying patient capital for companies and infrastructure projects whose requirements exceed grant-supported development. No allocation, launch timetable or selection framework has yet been formally approved.
