India’s new investment announcements jumped sharply in the June quarter. This boom looks good only on paper. Power and data centres together accounted for almost three-fourths of all new project proposals. This hides a slowdown in much of the investment economy. Total new projects in April–June rose 30 per cent compared with the same period last year to about Rs 14.5 lakh crore, up from Rs 11.3 lakh crore a year earlier. Compared with the quarter, proposals increased by roughly 3 per cent. However, this growth was driven by two capital-heavy sectors.
Data-centre projects alone added Rs 4.5 lakh crore, led by mega plans such as a large data-centre project in Raigad valued at nearly Rs 2 lakh crore and an integrated green data-centre park and AI compute hub in the Mumbai region worth about Rs 1.1 lakh crore.
Nuclear power drew over Rs 6 lakh crore in investment, including major projects such as a proposed nuclear plant in Maharashtra and another in Ratnagiri, each pegged at about Rs 2 lakh crore. These numbers reflect two underlying themes. India needs more computing infrastructure as artificial intelligence workloads expand.
This is pushing companies to build hyperscale, AI- data centres. Also, nuclear energy has gained policy backing as a source of round-the-clock "baseload" power to support that compute demand and broader industrial growth.
The government has been trying to build an "AI stack" with an AI mission worth over Rs10,300 crore across five years and a compute portal offering access to tens of thousands of GPUs and more than a thousand TPUs.
Nuclear power has also moved up the policy agenda: installed capacity is 8.78 GW today with projections to reach 22.38 GW by 2031–32 and a longer-term goal of 100 GW by 2047 under a dedicated nuclear energy mission.
A recent law allows private participation in the nuclear sector under strict regulation. Budget measures have reinforced this push by extending customs duty exemptions on equipment imports for nuclear power projects until 2035 and broadening the benefits to all plants, regardless of size.
Outside these capital-intensive niches, the investment picture is much weaker. Manufacturing investment proposals dropped from about Rs 3.7 lakh crore in the quarter to Rs 2.04 lakh crore. This shrunk the sector’s share of investments from roughly one-third to just 14 per cent. Mining proposals fell by 90 per cent compared to the same period last year. Transport services projects declined from about Rs 3.3 lakh crore to Rs 0.26 lakh crore. If power and data centres are excluded, new investment proposals fell by almost 74 per cent compared to the previous quarter and by 67 per cent compared to the same period last year.
This points to a loss of momentum in core private investment. The data-centre boom itself extends beyond the projects captured in this quarter.
One major estate and infrastructure developer recently announced plans to scale data-centre capacity to 2–3 GW over the next five years as part of a USD35 billion investment programme.
The company currently runs 250 MW and is discussing three new projects that could push it past 1 GW. Such data-centre projects are complex. They require parcels of land, reliable high-capacity power, advanced cooling, fibre networks, chips, cloud customers and long-term energy contracts. Artificial intelligence-heavy workloads make these economics even larger and more demanding.
Funding Gap for Startups and Broader Private Investments
Despite the infrastructure pipeline, private funding for AI startups remains modest compared to these large numbers. A report by the Deep Tech Alliance and other trackers notes that while AI and deep tech funding increased in 2025, India still lags behind the US and China in total private AI funding. Most of the capital continues to flow into infrastructure and core models rather than a wide range of application-layer startups.
The latest Economic Survey shows that India’s generative AI ecosystem saw the number of active startups jump from 66 to over 890 in two years. Cumulative funding rose only from 606 million to 990 million dollars in the same period. This means there was a 13-fold rise in the number of startups but only a 1.6-fold growth in capital. Within generative AI, infrastructure and model development still accounted for about 63Per cent of funding in the first half of 2025.
A separate report notes that while deep-tech startups raised 1.23 billion dollars across 120 rounds in 2026, far fewer deals have fallen. There were 372 deals in 2025. Only 120 in 2026. This indicates larger investments rather than a broad-based surge.
Founders and investors are speaking out about this issue. At an AI funding review, an AI-focused investor said capital was "becoming more concentrated”. VCs are writing cheques into fewer companies. They are also scrutinising whether startups have a tech advantage and a path to commercialisation. |
