Building capacity: Data centres in India to improve AI capabilities

Alphabet’s decision to build a mega artificial intelligence (AI) data hub in Visakhapatnam could be transformational for India’s fast-growing data-centre industry. The commitment by Google’s parent to make a $15 billion investment in the project over five years comes at a time of high tensions between India and the United States (US) due to the tariffs imposed by the latter. The project would be the largest data hub outside the US for Google, which has data centres in 12 countries. The investment ignores geopolitical tensions and potential legal complications, implying that the Alphabet management reckons the potential rewards outweigh the risks. Despite the legal issues, Alphabet sees India as a key market. It is a large market for Android phones and also has the world’s largest YouTube audience. The planned hub could create 188,000 jobs. The Adani group and Bharti Airtel will partner Google to build the infrastructure, including laying a new international subsea gateway. The project will combine Cloud and AI infrastructure powered by renewable energy systems and connected through a fibre-optic network. It is part of the Andhra Pradesh government’s plan to develop 6 Gw of data centre capacity by 2029. 

These centres are physical facilities housing the computing and networking equipment required to collect, process, store, and distribute data. Apart from real estate, enormous power is required to run the systems and keep things cool. This is why data-centre capacities are measured in gigawatts (a unit of power) rather than gigabytes. Moreover, that power must be reliable and available 24x7 without voltage fluctuations, which makes this a tricky engineering challenge. The state government has offered subsidised land and power to attract investors. It is one of the prime movers among states aiming to cash in on the policy changes that have driven growth in the data-centre industry. Digital infrastructure is crucial for creating the physical capacity to build a $1 trillion digital economy, as targeted by 2027-28. 

India has other advantages too. Data is cheap, and India is the world’s largest data consumer, with a large, growing, digital population. But along with physical capacity and the ecosystem, the legislative and regulatory policy changes are crucial if global players are to be persuaded to locate data centres here. Multinationals must believe that their data is safe and secure in servers located within India. India’s policy pushes for data localisation and security through initiatives like the Digital Personal Data Protection Act and the accompanying rules. The commitment of Google and those earlier by the likes of Microsoft and Amazon validate the policy. Data-centre capacities are coming up in Mumbai, Chennai, Pune, Hyderabad, the National Capital Region, Bengaluru, and Kolkata, as many states are trying to attract investment. India’s data-centre capacity crossed 1 Gw in 2024, up 200 per cent since 2019. Apart from employment, the location of this mega project in India should help to accelerate India’s own AI mission. AI requires enormous computing power, which is pushing the demand for specialised centres where thousands of chips can be clustered together in custom-designed servers. The learning from Vizag and other such centres should help India develop the capacity it needs to create and exploit advancements in computational capabilities.