India weighs Rs 20,000-crore AI fund. What will the money actually build?

India is considering a Rs 15,000-20,000 crore fund to finance advanced artificial intelligence (AI) companies and expand the computing infrastructure needed to develop and run AI models, as the government looks to address a key bottleneck in the country’s AI ambitions, The Economic Times reported on Friday.

The proposed National Frontier AI & Compute Fund (NFAICF), being explored under the IndiaAI Mission, could provide long-term capital to domestic AI startups and support the development of graphics processing unit (GPU) clusters, specialised data centres and other computing infrastructure.

The proposal is still under discussion, with the fund’s size, investment structure, governance and the government’s role yet to be finalised, the report said.

The initiative comes as the cost of training advanced AI models and building the infrastructure needed to support them rises, making it harder for Indian startups to compete for capital through conventional venture funding and government grants.

What will the proposed AI fund finance?

The proposed fund is expected to focus on two areas, financing companies developing advanced AI models and building the computing infrastructure required to support them.

On the technology side, it could provide long-term risk capital to Indian companies working on frontier AI models. On the infrastructure side, the fund could support GPU clusters and specialised data centres that provide the computing power needed to train and run these models.

More from Business

The investment is aimed at projects that require substantial upfront capital and may not find adequate funding through traditional venture capital channels.

The scale of the challenge is growing. According to a study by research institute Epoch AI, the cost of training large frontier AI models could exceed $1 billion per run by 2027.

The proposed fund could help bridge the financing gap by bringing government-backed capital into projects that require large, long-term investments.

How will the government and private investors share the cost?

The fund is being designed to attract private investment, with the government potentially acting as an anchor investor to encourage capital flows into India’s frontier AI ecosystem.

K Bhaskar Varma, chief business officer at NxtGen Datacenter & Cloud Technologies, said discussions appeared to be exploring a funding model in which the government and private investors could contribute equally.

“The intent appears to be to explore a 50:50 model with private participation,” Varma said, pointing to the high costs involved in building large-scale AI infrastructure.

The discussions have also considered whether international institutions such as the International Finance Corporation could participate in the proposed fund.

However, the final funding arrangement and the extent of government participation remain undecided.

How is it different from the IndiaAI Mission?

The proposed fund would complement the government’s existing IndiaAI Mission, which has an approved outlay of Rs 10,372 crore to develop India’s AI ecosystem.

The mission supports access to computing capacity at subsidised rates and provides funding for indigenous foundation models, alongside other initiatives aimed at strengthening the country’s AI capabilities.

The proposed NFAICF, in contrast, is being considered as a dedicated financing mechanism for frontier AI companies and large-scale computing infrastructure.

The government has identified a gap in domestic equity funding for companies developing advanced AI technologies. The new fund could address this by providing long-term capital to businesses and infrastructure projects that require substantial investment.