AI-led data centre expansion gathers pace in India amid power, policy challenges

India is set to see a sharp acceleration in data centre investments over the next two years as global demand for artificial intelligence (AI), cloud computing and digital services drives a worldwide surge in capacity, according to a new outlook by Moody’s Ratings published on Tuesday. While the surge presents a significant opportunity for India, it also underscores mounting challenges around power availability, regulatory clarity and rising costs.

Globally, data centre electricity consumption is projected to touch about 600 terawatt-hours (TWh) in 2026, up 14% from 2025 and nearly 40% higher than 2024 levels. Most of this capacity is being built to support AI training, cloud workloads and digital services, with large technology companies pre-leasing facilities to secure faster access to computing power.

Asia-Pacific is expected to be among the fastest-growing regions, and India is emerging as a key beneficiary alongside China and Southeast Asia, driven by lower construction costs, a large digital user base and growing cloud adoption.

Hyperscalers anchor demand

The bulk of new capacity is being locked in by global hyperscalers such as Amazon Web Services, Microsoft, Google and Oracle. Pre-leasing reduces the risk of unoccupied capacity for developers but increases concentration risk, as revenues depend heavily on a small group of tenants.

This model is already visible in India. Hyperscalers are expanding cloud regions and availability zones, while domestic and international data centre operators—including Sify, Yotta, CtrlS, AdaniConneX, ST Telemedia Global Data Centres, Equinix and NTT to name a few—are scaling facilities across Mumbai, Chennai, Hyderabad and Noida. Much of this capacity is designed to support AI workloads, enterprise cloud migration and data localisation requirements.

Moody’s estimates that Asia-Pacific will require up to $900 billion in investments by 2030 to meet data centre demand, with India contributing a meaningful share of incremental power and infrastructure needs.

Power remains the biggest constraint

Despite the growth opportunity, access to reliable and affordable power remains the biggest bottleneck. Data centres require uninterrupted electricity, making grid connectivity, redundancy and long-term power contracts critical.

“For large markets like India, incremental power demand from data centres will be added to already ambitious power development pipelines aimed at supporting broader economic growth,” Moody’s noted.

India has made rapid progress in renewable energy, but data centres continue to rely heavily on thermal power to ensure reliability. In established hubs such as Mumbai and Chennai, developers face delays in securing grid connections, while water usage and land availability have emerged as local concerns. Some operators are exploring on-site power generation and captive renewable arrangements to bridge gaps, but these raise costs and regulatory complexity.

Faster builds, higher costs

Globally, hyperscalers are prioritising speed to market, pushing developers to compress construction timelines. To accelerate delivery, tenants are increasingly willing to share construction risks—such as exemptions related to power or utilities availability at completion—something rarely seen earlier.

At the same time, construction costs are rising due to shortages of skilled labour, cooling systems, copper, rare earth metals and advanced power equipment. For Indian developers, this means newer facilities are more expensive to build than older ones, leading to higher lease rates even as demand remains strong.

Industry executives say this dynamic is already visible in India’s primary markets, where premium AI-ready data centre capacity commands higher pricing due to power density, cooling requirements and redundancy standards.

Financing shifts as GPUs dominate capex

Another trend with growing relevance for India is the soaring cost of GPUs, which now represent the single largest capital expense in AI-focused data centres—often exceeding the cost of the facility itself. As a result, developers and tenants are exploring alternative financing structures beyond traditional bank loans.

Moody’s expects greater participation from institutional investors, private credit funds and infrastructure investors as capital requirements rise. This could open opportunities for Indian infrastructure platforms and global funds to back long-term, pre-leased hyperscale projects, particularly those anchored by creditworthy tenants.

Policy clarity will shape outcomes

While countries such as the US and Brazil are adjusting regulations to attract AI infrastructure investment, Moody’s cautions that regulatory uncertainty can slow execution. In India, clear policies around power tariffs, water usage, environmental approvals and AI governance will be essential to sustaining momentum.

With global data centre investments projected to exceed $3 trillion over the next five years, India is well-positioned to ride the AI-led infrastructure boom. But the pace at which it can scale will depend on how quickly power constraints are resolved and policies aligned with hyperscalers’ increasingly urgent timelines.