AirTrunk commits $30B to build 5GW of AI data center capacity in India
A 5GW capacity commitment signals India's rapid transition into a primary global AI compute destination. Delivering this scale will require unprecedented grid modernization, dedicated renewable energy pipelines, and advanced liquid cooling deployments to handle high-TDP AI accelerators. This megaproject will fundamentally strain and alter the regional supply chain for high-density power equipment.
AirTrunk’s announcement of a $30 billion investment to build 5 gigawatts (GW) of AI-focused data center capacity in India is a watershed moment for global compute infrastructure. This massive capital deployment aims to position India as a primary hub for AI training and inference, moving beyond its traditional role as an IT services provider.
Technical Implications From an engineering standpoint, 5GW of power is staggering—roughly equivalent to the power draw of a major metropolitan city. Designing facilities for modern AI workloads requires planning for ultra-high rack densities. While traditional cloud data centers operate at 10-15 kW per rack, AI clusters utilizing next-generation silicon push rack densities to 100-120+ kW. To support this at a gigawatt scale, AirTrunk will need to implement advanced direct-to-chip (D2C) liquid cooling and rear-door heat exchangers as baseline standards, bypassing legacy forced-air systems. Furthermore, securing 5GW of reliable, continuous power in India will necessitate massive dedicated substations and direct integration with utility-scale renewable energy projects to satisfy both technical uptime requirements and ESG mandates.
Why It Matters This investment highlights a strategic geographic diversification of AI compute. As power grids in traditional hubs like Northern Virginia, Silicon Valley, and Ireland face severe capacity constraints and multi-year waitlists for new deployments, hyperscalers are forced to look toward emerging markets with aggressive infrastructure growth. India offers the land, a rapidly growing domestic AI market, and the governmental push for digital infrastructure necessary to absorb this scale of demand.
What to Watch Next The primary bottleneck for this megaproject will not be capital, but physics and supply chain realities. Watch for AirTrunk's strategy regarding Power Purchase Agreements (PPAs) with major Indian energy conglomerates to secure the actual generation capacity required. Additionally, monitor their procurement pipeline for high-voltage transformers, switchgear, and coolant distribution units (CDUs), as global lead times for these critical-path components currently stretch beyond 18 to 24 months.