On the morning of April 28, excavators began moving across 601 acres of farmland outside Visakhapatnam, the east-coast Indian city better known as Vizag.
The project is not a routine industrial park. It is Google's largest commitment in India to date: $15 billion over five years, three data center campuses and a plan to create the country's first gigawatt-scale AI base. India's railway minister, Andhra Pradesh chief minister N. Chandrababu Naidu, senior Adani Group figures and Bharti vice chairman Rajan Mittal all attended the groundbreaking, giving it the feel of a national infrastructure launch rather than a private tech project.
This Is Not a Normal Data Center
The headline numbers are large even by AI infrastructure standards: $15 billion of investment from 2026 to 2030, 601 acres across Tarluvada, Adavivaram and Rambilli, three data center campuses, gigawatt-class power demand and an expected 200,000 direct and indirect jobs. The project is scheduled to be unveiled in September 2028.
Gigawatt-scale matters. Conventional data center campuses usually sit in the tens or hundreds of megawatts. A gigawatt, or more than 1,000 megawatts, is closer to the electricity draw of a mid-sized city, and it has to be supplied continuously and reliably. India has not previously had AI infrastructure of this scale dedicated to inference and cloud workloads.
Why Google Chose Adani and Airtel
The more important detail is the partnership model. Google is not building the entire stack alone. AdaniConneX, part of Adani Group, is responsible for the main data center campuses and renewable power supply. Nxtra by Airtel, Bharti Airtel's data center arm, is handling high-speed fiber, submarine cable access and operating support. Google Cloud will bring the workloads, including Gemini, Google Search and AI inference.
The Adani role is central because power has long been the weak point for data center operators in India. Grid stability, outage frequency and industrial electricity costs are all difficult variables. Adani has renewable energy assets across solar and wind, plus experience in ports and industrial parks. Giving that side of the project to Adani is a way to reduce the power-risk problem at the source.
Airtel's role is connectivity. Vizag is on India's east coast, with access to submarine cable systems running toward Singapore and Europe. AI inference is latency-sensitive, so the location effectively ties an Indian AI compute hub directly into the global internet backbone.
Why Vizag, and Why Now
The site choice is deliberate. Andhra Pradesh has become one of India's most aggressive state governments in courting technology investment, with Naidu making frequent trips to Silicon Valley. The exact state incentives have not been disclosed, but the land assembly, unified planning and coordination around village relocation point to a heavy public-sector push.
The timing also fits the market. AI inference demand in India has grown quickly over the past 12 months. Gemini is widely available in the country, AI Overviews has reached Google Search users in India, and many domestic AI startups already run on Google Cloud. Until now, much of Google's India-facing AI compute has had to route through Singapore or Japan, adding at least about 30 milliseconds of latency.
"Defining Milestone For India's AI Journey"
-- Thomas Kurian, CEO of Google Cloud
Kurian's line is more than launch-stage phrasing. India is Google Cloud's largest Asia-Pacific market outside Japan, but a meaningful share of Indian workloads has historically run from regions such as Singapore or Taiwan. Moving gigawatt-scale compute into India addresses latency, compliance and data-sovereignty pressure at the same time.
An Underrated Signal
The political signal is just as important. The Trump administration's stance toward India's technology sector has been tense this year, with friction around visas, tariffs and export controls. Even so, the $15 billion project moved ahead and was presented under India's Viksit Bharat development agenda.
The logic is straightforward: AI infrastructure is no longer something Big Tech can carry alone. It needs countries with real power, available land and large domestic demand. U.S. power supply is constrained, Europe is heavily regulated and the Middle East faces water and power trade-offs. India, long underweighted in AI infrastructure discussions, now has population, electricity ambition and geopolitics lined up in its favor.
The next tests are whether Andhra Pradesh can deliver its power commitments by 2028, whether Adani's renewable buildout keeps pace, and whether Indian AI companies actually move workloads back into local Google Cloud regions. If those pieces hold, India's AI infrastructure could redraw compute distribution across Asia-Pacific, pulling some demand south from Singapore, Korea and Japan. The real verdict on those 601 acres in Vizag will come in September 2028.
Sources: Google Breaks Ground on India AI Hub, Launching a National Industrial Ecosystem (Google Cloud Press Corner); Google begins work on $15 bn AI hub in Vizag (Business Standard); Visakhapatnam set to become global data gateway (The Week); CocoLoop; Defining Milestone For India's AI Journey: Says Google Cloud CEO Thomas Kurian (Free Press Journal)