AirTrunk commits $30B to build 5GW of AI data centers in India by 2030

Blackstone-backed data center operator AirTrunk has committed $30 billion to develop 5 gigawatts of new data center capacity in India by 2030, one of the largest single infrastructure pledges directed at AI compute in the country's history.

What's new

The commitment centers on a flagship 3GW campus at the Raigad Pen Growth Center in Maharashtra, representing roughly $21 billion of the total investment. AirTrunk already operates a 600MW pipeline across Mumbai, Chennai, and Hyderabad, bringing its India-wide buildout to more than 5.6GW when fully built.

The announcement coincides with a series of policy moves by the Indian government designed to attract foreign AI infrastructure investment. New Delhi is now offering foreign cloud providers tax exemptions through 2047 on services sold overseas, provided those workloads are run from Indian data centers — a direct incentive for hyperscalers to co-locate with AirTrunk rather than build independently.

India's total data center capacity currently stands at roughly 1.5GW; analysts project it reaching 8GW by 2030, with AirTrunk's build-out accounting for more than half of the anticipated new supply.

Context

AirTrunk, headquartered in Australia, was acquired by Blackstone in 2024 for approximately $16 billion — at the time the largest data center deal in Asia-Pacific history. Since then, the firm has aggressively expanded its Asia-Pacific footprint under Blackstone's ownership, with India now becoming its largest single-country commitment.

The India market has emerged as a priority for AI infrastructure investment in 2026. Major cloud providers including Google, Microsoft, and Amazon have each announced multi-billion-dollar India commitments over the past 18 months, drawn by a combination of regulatory incentives, a large English-language developer base, and rapidly growing enterprise AI adoption.

Data center development at this scale carries significant resource constraints. Power, water, and land availability remain potential bottlenecks, particularly for the Maharashtra campus — a point AirTrunk has acknowledged publicly as a key execution challenge.

Why it matters

A 5GW commitment is roughly comparable in scale to the full AI data center capacity the United States added in 2023. For the AI supply chain, India moving from 1.5GW to a projected 8GW by 2030 represents a structural shift in where frontier AI inference — and eventually training — capacity is located.

For hyperscalers and AI labs deploying models at global scale, the combination of tax incentives and purpose-built colocation at this density creates a credible alternative to European and US-East deployments for serving South and Southeast Asian demand. The 2047 tax exemption timeline also signals long-duration policy stability rarely offered in emerging markets.

The announcement adds to a pattern of non-US infrastructure bets accelerating in 2026: the EU's AI Continent action plan, Gulf sovereign wealth commitments, and now India's subsidy architecture are collectively reshaping where AI compute power concentrates.