Adani Commits $100 Billion to Build India’s AI Data Center Empire

The Adani Group announced a $100 billion investment to build renewable-energy-powered, hyperscale AI data centers across India by 2035. The commitment, disclosed on the second day of the India AI Impact Summit in New Delhi, represents the single largest corporate pledge to AI infrastructure in India’s history and one of the biggest globally.

The investment targets 5 gigawatts of data center capacity—up from AdaniConnex’s existing 2 GW platform—through partnerships with Google and Microsoft across four major campuses. Adani estimates the direct spending will catalyze an additional $150 billion in server manufacturing, sovereign cloud platforms, and supporting industries, creating a $250 billion AI infrastructure ecosystem over the decade.

“Nations that master the symmetry between energy and compute will shape the next decade,” Chairman Gautam Adani said in the announcement.

Google and Microsoft Anchor the Build-Out

The data center expansion is anchored by two hyperscaler partnerships that give the project immediate credibility and demand.

With Google, Adani is building what will be India’s largest gigawatt-scale AI data center campus in Visakhapatnam, Andhra Pradesh, with additional facilities in Noida. Google committed $15 billion over five years to the Visakhapatnam hub, which will include new international subsea cables and be built to the same specifications powering Google Search and YouTube globally. The first phase alone targets roughly 1 GW of power capacity.

Microsoft’s partnership spans campuses in Hyderabad and Pune, adding geographic diversity to the network. Flipkart has also signed on for a second AI data center focused on digital commerce and high-performance computing.

The partnerships address one of the fundamental challenges in data center development: guaranteed demand. Building gigawatt-scale facilities is capital-intensive and risky without committed anchor tenants. Having two of the world’s largest cloud providers locked in de-risks the investment and signals to other potential customers that the infrastructure will meet global standards.

Adani isn’t just building the facilities—it’s also building the power supply to run them. The company’s 30 GW Khavda renewable energy project in western India, with roughly 7 GW operational at the site and over 11 GW across Adani Green Energy’s full portfolio, provides the energy foundation. The company says it plans significant additional investment to expand renewable generation and battery storage, including what it describes as one of the world’s largest single-location battery energy storage systems.

The renewable angle matters because energy costs represent the largest ongoing expense for data centers, and carbon-neutral power is increasingly a procurement requirement for hyperscalers like Google and Microsoft. By controlling both the data center and the energy supply, Adani can offer competitively priced, clean compute—a combination that few global competitors can match at this scale.

India’s Bid for AI Sovereignty

The announcement arrives at a moment when India is aggressively positioning itself in the global AI race. The India AI Impact Summit—the first global AI summit hosted in the Global South—has drawn heads of state from roughly 20 countries alongside CEOs from Meta, Microsoft, Google, OpenAI, and Nvidia. India’s government AI Mission has committed approximately $1.2 billion to building domestic AI capabilities.

The country’s AI adoption is already accelerating. India is now home to 100 million weekly ChatGPT users, making it OpenAI’s second-largest market. But consumption alone doesn’t build an AI economy. What India has lacked is the physical infrastructure layer—the data centers, chips, and energy systems—needed to train and serve AI models domestically rather than relying on compute capacity located in the U.S. or Singapore.

Adani’s plan explicitly targets this gap through what the company calls a five-layer sovereign AI stack: applications, models, chips, energy, and data centers. The company is reserving GPU capacity for Indian startups, research institutions, and deep-tech entrepreneurs—a detail that distinguishes this from a pure commercial real estate play. It’s also developing specialized AI infrastructure engineering curricula with Indian universities and localizing supply chains for transformers, power electronics, and thermal management systems.

The sovereign AI framing is deliberate. Countries from Saudi Arabia to France have launched national AI infrastructure programs, recognizing that dependence on foreign compute creates strategic vulnerability. India’s approach differs in scale and structure: rather than government-built facilities, it’s leveraging a private conglomerate with existing energy and logistics assets to build infrastructure that serves both commercial and national interests.

The Scale in Context

At $100 billion, Adani’s commitment dwarfs most individual AI infrastructure spending announcements. For comparison, Amazon plans to spend $200 billion on AI and cloud infrastructure globally over the coming years, and Microsoft has committed $80 billion for fiscal 2025 alone. But those figures span worldwide operations. Adani’s pledge is concentrated entirely in one country.

The bet carries real risk. Adani Group faced a credibility crisis in 2023 after a short-seller report wiped roughly $150 billion from its market value, though the conglomerate has since recovered significantly. Executing a decade-long, $100 billion infrastructure program requires sustained access to capital markets, stable regulatory conditions, and continued demand growth for AI compute.

But the debate over infrastructure versus innovation in AI may be converging. As AI companies command valuations in the hundreds of billions, the physical layer—power, cooling, connectivity—has become the binding constraint. Adani is betting that whoever controls that layer in the world’s most populous country controls a critical piece of the global AI economy. Given the hyperscaler partnerships already in place, that bet looks less speculative than it might have a year ago.