The government is considering a ₹15,000–20,000 Cr anchor investment in a proposed National Frontier AI & Compute Fund to finance frontier AI startups, GPU clusters and specialised data centres
The proposed fund could bring long-term private capital into a capital-intensive segment where model development and compute infrastructure require significantly larger investments
The proposed fund could be structured as a SEBI-regulated Category I AIF, with a potential 50:50 government-private capital model
The Centre is considering an anchor investment of ₹15,000 Cr to ₹20,000 Cr in a proposed fund aimed at supporting frontier AI companies and expanding the country’s compute infrastructure.
The proposed National Frontier AI & Compute Fund (NFAICF) could provide long-term risk capital to AI companies and finance GPU clusters, specialised data centres, and other infrastructure required to develop and deploy advanced AI models, ET reported, citing sources.
The proposal is currently under consultation, and the fund’s final corpus, structure, governance framework, and the role of government capital are yet to be decided. It is also unclear whether the government would invest directly in companies and infrastructure projects or route the capital through an intermediary structure.
The government is considering a model under which it and private investors would contribute capital in a 50:50 ratio, the report said.
The proposed fund was discussed at a closed-door meeting on “Financing India’s Frontier AI & Compute Ecosystem” in Delhi yesterday. The meeting was attended by Sarvam cofounder Vivek Raghavan, Peak XV Partners MD Rajan Anandan, E2E Networks CEO Tarun Dua, MeitY secretary S Krishnan, and officials from the IndiaAI Mission.
The government sought views from VC firms, infrastructure providers, and other stakeholders on its role in attracting private capital to the frontier AI ecosystem.
The NFAICF could be structured as a SEBI-regulated Category I AIF, with investment decisions taken by a committee comprising experts from the technology, investment, infrastructure, and finance sectors.
The fund could invest through venture and growth equity, convertible instruments, infrastructure equity, fund-of-funds commitments, and direct co-investments, depending on the requirements of a company or infrastructure project.
Its proposed objectives include providing long-term capital to foundation-model builders, specialised AI systems and other strategically important AI technologies, while also financing GPU clusters, specialised data-centre capacity and other compute infrastructure.
The government is also examining whether international institutions such as the International Finance Corporation could participate in the proposed fund.
A key focus of the discussions yesterday was financing GPU infrastructure without creating excess or underutilised capacity.
The government is considering linking compute investments to demand from AI companies and research institutions through capacity commitments, utilisation arrangements or pre-agreed commercial terms. The objective is to reduce the risk of stranded or underutilised infrastructure.
It is additionally considering a follow-on and co-investment capability, allowing the proposed fund to participate in later funding rounds alongside private investors when strategically important companies require additional capital.
The final design of the NFAICF will depend on stakeholder feedback and further assessment.
This comes as the demand for advanced GPUs continues to outpace supply, with infrastructure providers facing longer delivery timelines and rising pressure to secure capacity. The capital required to build GPU infrastructure is also challenges for Indian cloud and neocloud providers, particularly as they look to expand their capacity for AI workloads.
The NFAICF is being considered in addition to the existing ₹10,372 Cr IndiaAI Mission, which provides subsidised compute access and funding for indigenous foundation models. The government has also been expanding the country’s AI compute capacity through the IndiaAI Mission.
The Mission initially targeted access to more than 10,000 GPUs through a public-private partnership to make high-end computing resources available to Indian AI companies and researchers. The government has since onboarded more than 38,000 GPUs through empanelled cloud service providers and plans to expand the compute pool further.
IndiaAI has also been supporting multiple companies and research teams to develop Indian foundation models across large multimodal models and smaller language models.
Source: Inc42
