Summary: On June 5, 2026, AirTrunk announced a $30 billion plan to add 5GW of India data-center capacity by 2030, including a planned 3GW Raigad campus. The article treats stable and renewable power as the decisive constraint because India’s grid expansion may not keep pace with projected AI demand.
June 5, 2026 was the day AirTrunk — the Australian data center operator with Blackstone and Canada’s CPP Investments behind it — went public with a plan: $30 billion invested in India by 2030, yielding 5 gigawatts of brand-new capacity. TechCrunch ranks it among the largest commitments India’s digital infrastructure has ever received. A reminder worth making: the company only set foot in India earlier this year, by way of its acquisition of Lumina CloudInfra.
Size is what makes this story. India’s data center capacity currently sits near 1.5GW, and Bernstein’s research projects the market could reach 8GW by 2030. Against that backdrop, one operator committing 5GW amounts to underwriting the country’s whole growth curve at corporate scale. Ahead of the announcement, CEO Robin Khuda paid a visit to Prime Minister Narendra Modi, who said in a post on X that the investment would “help strengthen India’s position as a global hub for cloud computing and artificial intelligence.”
Inside the $30B: A 3GW Campus at Raigad, Plus a 600MW Pipeline
The biggest project named so far is in Maharashtra. Chief Minister Devendra Fadnavis announced that the state and AirTrunk have exchanged letters of intent over land at the Raigad Pen Growth Center, earmarked for a 3GW data center campus carrying roughly ₹2 trillion (about $21 billion) in investment. One campus, in other words, soaks up about two-thirds of the headline figure.
The company’s existing presence in India amounts to a development pipeline of about 600MW spread across Mumbai, Chennai, and Hyderabad. Whether Raigad makes up the bulk of the 5GW — and whether other states will host more projects — is something AirTrunk won’t say. Mint, drawing on Reuters, reports the pipeline will span multiple states and support India’s push to become a global hub for AI and cloud investment.
The Case for India: Tax Breaks Through 2047 and Demand Headed for 8GW
Two forces make the wager defensible. Policy comes first: earlier this year, New Delhi laid out a scheme under which foreign cloud providers pay no tax through 2047 on services sold overseas, so long as the workloads run from data centers inside India — a rule shaped for precisely this kind of bet. Demand comes second: Bernstein expects capacity to climb from roughly 1.5GW to a possible 8GW by 2030, with AI inference doing most of the pulling.
Khuda lays out the investment thesis on three supports: backing from government, a deep bench of technical talent, and the ability to reach renewable energy. All three map onto India’s greatest strength and its greatest vulnerability.
The Crowd: No Giant Is Sitting Out
AirTrunk has plenty of company in India. Going by the rundown TechCrunch compiled, essentially every heavyweight has put money on the table there over the past year and a half:
- Amazon: another $35 billion by 2030, announced December 2025, stacked on top of $75 billion in total planned investment
- Google: a $15 billion AI infrastructure hub, announced October 2025
- Microsoft: $17.5 billion by 2029
- OpenAI: picked up 100MW from Tata in February 2026, and is reportedly eyeing 1GW
- Reliance Industries carries a $110 billion AI investment plan; Adani Group has pledged $100 billion for AI data centers
Measured against that field, AirTrunk’s 5GW is the biggest single-operator commitment to date — with the catch that land, grid capacity, and engineers must now be contested against rivals of equal financial firepower.
Electricity Is the Real Choke Point
Data centers run on electricity, water, and land in equal measure, and the analysts and executives commenting on this boom converge on power as the binding constraint. Deloitte reckons Asia-Pacific’s data center build-out could need tens of additional terawatt-hours of electricity by around the end of the decade. Read that way, Khuda’s choice to fold renewable energy access into the thesis looks less like stock ESG language and more like hedging: how much of the 5GW blueprint actually gets built will turn on securing power that is both stable and clean.
That is where the India story strains against itself. Policy and demand line up; the grid’s pace of expansion may not match an 8GW-sized ambition.
Takeaways for Developers and Product Teams
Three concrete consequences. One: AI products aimed at South Asia gain a nearby inference option, with 5GW of in-country capacity addressing latency and data residency head-on. Two: the tax exemption running to 2047 turns “India as an export base” into a formal playbook — global teams can shortlist the country when siting regional inference hubs. Three: the capacity market keeps deepening, with hyperscalers and independent operators alike opening more routes to large-scale inference capacity and more leverage in negotiations. Our 2026 opening outlook made the case that compute access would be the pivotal variable in product decisions this year — India is the newest and biggest footnote to that trend.