By Anjani Kumar
India’s data centre capacity is expected to scale from ~1.5 GW in 2025 to ~10- by 2030 as per the Deloitte – Niti Aayog report.
That scale demands compute availability, data centre commissioning velocity, and reliable, affordable power at scale. India’s advantage is that much of the AI era buildout is still ahead, so standards and enabling infrastructure can be designed before constraints harden.
Budget policy is therefore material because if power, approvals, and finance are not de-risked, delays and higher delivered tariffs will compress project IRR and slow supply as AI demand accelerates. Thus, the budget could focus on multiple areas as highlighted below.
AI data centres need connectivity, zones and fast clearances
The first focus area should be power grid and substation readiness because most of the data centre capacity is concentrated in a few city hubs (Mumbai, Chennai). As per the Deloitte – NITI Aayog report, India will need an additional 40-45 TWh of power capacity by 2030 to cater to the growth in data centre capacity.
India can mitigate this by building power-ready dedicated data centre zones (DCZs) ahead of demand, with pre-built substations, standardized connection timelines, and bankable firm power for high-density campuses.
It should also aim to standardise state-level renewable energy banking regulations and charges, since current variations create operational uncertainty for integrating renewables sources.
City connectivity determines where high-density AI campuses can run at low latency. AI grade facilities typically need ~100–400 Gbps bandwidth. Several potential DC hubs still lack low-latency fibre backhaul and leased high-speed lines remain expensive.
Budget support can shift connectivity from a site-by-site negotiation to a planned utility-like backbone. Targeted public capex can fund metro and intercity fibre connectivity, to designated data centre zones.
Sustainable growth needs green power, skills and execution
Commissioning speed is the third lever because infrastructure delays convert directly into cost. Setting up a data centre can require 40+ approvals across agencies and jurisdictions, and thus approvals should move to a single-window clearance system with fast, time-bound decisions.
The Draft National Data Centre Policy 2025 re-iterates the creation of Data Centre Economic Zones (DCEZ) through a central framework and state-led implementation. Some states are already adopting such zone-style constructs. A formal national notification of the DCEZ construct would standardise adoption across states and speed up the rollout.
Budget support for DCEZs can enable plug-and-play clusters with pre-built utilities and predictable timelines, reducing delivery risk for each additional campus. These zones should be designed for AI-era high-density compute and should be provisioned with access to firm power and industrial water.
As per industry estimates, India’s data centre sector could attract ~USD 30 billion of colocation facility construction capex by 2030. This scale-up will increase pressure on local power and water utilities, so new capacity must be matched with renewable power and water-sustainable sourcing.
Large buyers, including international customers, increasingly screen credible energy and water sustainability in vendor selection.
India also has an efficiency gap, with average PUE of ~1.9 versus ~1.3 for more efficient designs, alongside with continued reliance on conventional energy at ~55% of total usage, and rising water stress in metro hubs such as Mumbai and Bengaluru.
Budget support can link incentives to PUE and WUE outcomes to accelerate efficient cooling, renewable procurement, and water-efficient operations. Talent and execution capacity must scale with infrastructure, with AI talent demand rising from about 600,000–650,000 to more than 1.25 million people by 2027.
Budget support should fund workforce development for data centre operations and core engineering roles. Public–private partnerships with educational institutions and technology firms can be used to develop specialised training programmes that focus on practical skills needed in the sector.
India now has early proof of a coordinated builder stack: a funded mission at ₹10,371.92 crore, a published compute price which is discounted versus global benchmarks, and measurable GPU installation numbers.
However, India can convert this into a global data centre hub only if the execution levers are delivered at scale to attract and retain investment. These include plug-and-play data centre zones, grid and substation readiness, backbone connectivity, and faster commissioning through time-bound clearances. If these are addressed, India can become a reliable, cost-competitive destination for global AI capacity build-out.
Anjani Kumar is a Partner at Deloitte India
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.