Highlights:
- Google to invest $15 billion in a new AI data centre hub in Andhra Pradesh, India.
- Project will be Google’s largest AI hub outside the US, built over five years.
- Google Cloud’s Karan Bajwa says AI is not a bubble but a deep investment.
- Over 60% of data centre costs come from GPU chips used for AI processing.
- Growing energy and water use may push more AI hubs to India and Asia.
Google is investing $15 billion to build a massive new data centre hub in Andhra Pradesh, India, as part of its global artificial intelligence (AI) expansion. The company said this project would become Google's largest AI facility outside the US and will be completed over the next five years.
At a media event in New Delhi, Google Cloud Asia Pacific president Karan Bajwa said that AI is not a short-lived bubble but a long-term statement. “If you look at the scale of money being spent to build AI infrastructure, it’s not just hype. We are building capabilities, bringing them to market, and customers are using them,” Bajwa said.
The event, called Bharat AI Shakti, was attended by finance minister Nirmala Sitharaman, IT minister Ashwin Vaishnaw, Andhra Pradesh chief minister N. Chandrababu Naidu, and Google Cloud CEO Thomas Kurian. The project highlights India's growing role in the global AI ecosystem and Google's confidence in the country's digital potential.
Companies like Google, Microsoft, and Amazon are racing to build more AI data centres to support the growing use of generative AI tools by businesses and consumers. But this expansion comes with high costs.
Data centres require thousands of GPUs (graphical processing units) and huge amounts of electricity and water to run and cool the systems. Nearly 60 per cent of the total cost of a data centre comes from GPU chips alone.
Some investors have started worrying about whether such massive spending will pay off soon. However, Bajwa believes the results are already visible. “AI is making a real difference in people’s lives,” he said. Another Google executive added that every time a new AI model is launched, it processes trillions of tokens within hours, showing how fast users adopt these technologies.
Still, environmental and energy concerns are rising. Many local communities in the US are opposing new data centres because of their heavy power and water use. Experts believe this could push more AI infrastructure projects to countries like India, where energy costs are lower and the digital ecosystem is expanding.
According to a McKinsey report, the demand for data centres in the US could triple by 2030, making them responsible for over 14 per cent of the country’s total power use.
