AirTrunk’s $30 billion India plan shows how quickly AI infrastructure has become a private capital race. The question now is not whether India needs more compute, but who will own the rails it runs on.
AirTrunk is putting one of the loudest numbers yet on India’s AI buildout: $30 billion for 5 gigawatts of data centre capacity by 2030. That is not a routine market entry. It is a statement that India is moving from promising digital economy to contested infrastructure prize.
According to Reuters, the Sydney-based data centre operator said on Friday that it would invest the money over the next four years, after founder and chief executive Robin Khuda met Prime Minister Narendra Modi in New Delhi. Modi framed the commitment as a boost for cloud computing and artificial intelligence, and that is exactly how investors will read it. AI demand is no longer only about chips, models and software teams. It is about land, power, cooling, debt, permits and tenants who can sign long contracts before the concrete is poured.
That is why this matters beyond India. AirTrunk is backed by Blackstone and CPP Investments, and Blackstone’s 2024 acquisition valued the business at more than A$24 billion, roughly $16 billion at the time. The firm was already one of Asia-Pacific’s strongest hyperscale platforms, with campuses across Australia, Japan, Malaysia, Hong Kong and Singapore. India now gives it a larger stage and a more complicated one.
India has the demand story investors like to underwrite. It is the world’s second-largest telecom market by subscriber base, with fast-growing cloud adoption, 5G expansion, tighter data sovereignty expectations and a government that wants domestic AI capability. Those forces do not guarantee good returns, but they do create the kind of long-duration need that infrastructure investors understand.
AirTrunk did not arrive empty-handed. In April, it acquired Lumina CloudInfra, another Blackstone-backed platform with operations and development plans across Mumbai, Chennai and Hyderabad. That deal gave AirTrunk a planned portfolio of about 600 megawatts and up to $5 billion of future development potential. The new $30 billion commitment is the much larger version of the same bet: get scale early, secure the best sites and become the partner hyperscalers can trust when demand outruns local supply.
The Maharashtra angle is especially important. Earlier this week, the state government signed a letter of intent with AirTrunk for a data centre project in Raigad district, near Mumbai, worth about 2 trillion rupees, or roughly $21 billion. Data Center Dynamics reported that the project could provide 3 gigawatts of IT capacity when complete. The letter is preliminary, but it shows where the first big fight is likely to happen. Mumbai is already one of India’s most important data centre clusters, and the next wave needs more land and more power than the city itself can easily provide.
Private equity is moving up the stack
For years, the technology story was told from the top of the stack. Startups built apps. Cloud platforms sold capacity. Chipmakers supplied the engines. Now the base layer is getting more attention because AI workloads are exposing the limits of everything underneath. The owners of data centres, substations, transmission agreements and renewable power contracts are no longer background players.
Blackstone has been preparing for this moment. Its AirTrunk deal followed years of investment in digital infrastructure, including QTS and other data centre assets. In May, Reuters also reported that Google and Blackstone would form an AI cloud venture in the United States, with Blackstone committing an initial $5 billion in equity to bring 500 megawatts of capacity online in 2027. That gives the India announcement a broader context. Blackstone is not simply buying buildings. It is trying to assemble a global AI infrastructure machine.
That strategy makes sense, but it carries risks. India’s data centre buildout will depend on reliable electricity, water management, grid upgrades and state-level execution. A 5 gigawatt plan is impressive on paper. In practice, each campus becomes a local infrastructure negotiation. Communities will ask who gets the jobs, who bears the power burden and whether these facilities support domestic AI companies or mostly serve global cloud giants.
There is also the tenant question. AirTrunk has not publicly named the hyperscalers behind the demand, and that matters. Large data centre projects usually need anchor customers with strong credit and long commitments. Without them, the headline number can move faster than the actual capacity. With them, India could become one of the few markets where global cloud platforms, sovereign AI ambitions and private capital all converge at the same time.
For entrepreneurs, the lesson is straightforward. AI opportunity is not limited to building the next model or chatbot. The bigger value chain now includes energy software, cooling systems, construction technology, grid services, compliance tools and specialist operations for high-density compute. When $30 billion moves into one market, smaller companies should watch the gaps it creates around the main project.
AirTrunk’s India plan will not be judged by the announcement. It will be judged by sites powered, tenants signed and capacity delivered before 2030. If Blackstone and AirTrunk can execute, India becomes a serious AI infrastructure hub. If they struggle, it will remind everyone that compute is easy to demand and hard to build.
Also read: Lightricks splits itself in two as AI costs force a reset • Bitcoin now has a path into Fannie Mae-backed mortgages • AI has become the clearest signal in the 2026 tech layoff wave
