Google’s $15-billion India AI bet is about more than data centres: Preeti Lobana

Google’s India artificial intelligence (AI) push will extend beyond data centres to compute capacity, subsea connectivity, local research and language datasets, as the company looks to become a “co-architect” of the country’s AI-led transformation, said Preeti Lobana, Country Manager, Google India, in an interaction with Business Today.

The company’s $15-billion investment in a gigawatt-scale AI hub in Vizag is central to that effort. Lobana said the project should be viewed as foundational infrastructure, encompassing undersea cable links for low latency as well as compute capacity that Indian businesses, startups and institutions can build on.

Must read: Google trims headcount the soft way: Offering voluntary buyouts instead of layoffs

“The role and vision for Google is how do you provide the foundations,” she said, adding that the company also wants to bring its global research capabilities into the Indian context and develop solutions around local needs.

The planned infrastructure has raised questions about energy and water use. Residents, environmental groups, and citizen coalitions in Vizag have launched organised protests against massive, proposed AI data centre projects. Lobana said Google would pay for the power it consumes, including new substations and transmission lines required to support the data centre. She said the facility would be air-cooled and that Google had signed agreements for 125MW of solar power with ReNew and 125MW of wind and solar power with CleanMax. The company also has water-replenishment commitments, she said.

Must read: Google tests pulling the plug on AI Mode’s front-row seat in Search

During the conversation, Lobana also highlighted that Google’s investment decisions are guided by whether AI can benefit citizens, enterprises,and communities, rather than a select group of users. Healthcare, education, and agriculture are among the areas where the company sees scope for impact. Making AI tools available in Indian languages is critical, she said, citing the example of a teacher in Kerala using AI in Malayalam, rather than only English.

Through Project Vaani, Google has recorded 109 languages and made datasets available to the government’s Bhashini project, Lobana said. It is also making open-weight models such as Gemma available to developers, alongside Gemini.

Google has committed to upskilling 10,000 people in the Vizag area in AI skills and providing cloud and AI training to around 1,200 students and teachers. It also plans to put tools in the hands of around 1,000 fishermen, Lobana said.

On AI safety, she said Google follows a “bold and responsible AI” approach, with checks built into product development, model red-teaming and frameworks for frontier and agentic AI. Citing SynthID, its watermarking technology for identifying AI-generated content, Lobana called for shared industry standards, saying safety is a collective responsibility.