Google Cloud has announced that it is making a major change in its artificial intelligence (AI) operations in India. Indian companies have already been using Gemini and other Google AI services through the cloud, with their data staying within India. What is changing is where Google's latest AI models run.
In an exclusive interview with ET, Google Cloud chief executive Thomas Kurian said the company is now deploying its latest AI models on infrastructure located inside India, allowing enterprises to keep their data as well as AI processing within the country. The move points to a broader shift underway in global cloud computing, where the location of AI infrastructure is becoming as important as the AI models themselves.
Google's latest announcement suggests it is planting its AI flag in India, treating India not merely as a market but as a strategic AI hub.
From access to local hosting
Kurian said in his interview with ET during his ongoing India visit that the company is bringing its latest AI offerings "to serve from machines in India." Until now, an Indian company could access Gemini through Google Cloud, but the actual AI inference process may not always have occurred inside India. In many cases, requests could be processed in another Google Cloud region depending on how the service was deployed and where the model was available. The customer was in India, but the model itself was not necessarily running there.
Kurian's announcement changes that equation. Google is deploying its latest Gemini models and Gemini Enterprise, its platform for building AI agents, on infrastructure physically located in India. As he put it, the goal is to enable "local processing with both the models and the data being hosted in India." In practical terms, this means an Indian bank, telecom operator, healthcare provider or government department can not only keep its data in India, as it already happens, but also have AI requests processed in India rather than being routed overseas.
Why Google is doing this now
The immediate driver is the growing importance of digital sovereignty. Governments around the world are becoming more sensitive about not only where data resides but also where AI systems process information. Enterprises in regulated sectors increasingly want assurances that sensitive data never leaves national borders and that AI requests are not processed outside them. The debate has intensified as countries begin to view AI infrastructure as a strategic asset, much like telecommunications networks or energy infrastructure.
Also Read: View: Meta's cloud play will spice up Ambani-Adani rivalry
Kurian said Google is expanding its commitment to digital sovereignty and described the latest deployment as part of the company's "strict in-country machine learning process commitment that we made to India, and we are following through on it with the latest model".
The timing is also notable. AI models have become a key layer of digital infrastructure, and countries are seeking greater control over how these systems operate within their jurisdictions. For cloud providers, offering local AI processing is becoming a competitive necessity rather than a premium feature.
What exactly will be hosted in India?
Based on Kurian's comments, Google is not merely storing customer data in India. It is also deploying the AI models themselves on Indian infrastructure. This shift is important because data residency and AI residency are not the same thing. A company may store its databases inside India while still sending AI requests to servers in another country. Google's new architecture aims to allow both the data and the model inference layer to remain local.
The announcement covers Gemini Flash and Gemini Enterprise, Google's platform for building AI agents. Enterprises using Google's AI stack will be able to run AI workloads on infrastructure hosted within India while maintaining access to the same underlying technology available globally. This does not mean the models were developed or trained in India. Gemini remains a globally developed Google model. What changes is where the model is deployed for customer use.
India's growing importance in Google's cloud strategy
Kurian's comments also reveal how central India has become to Google's cloud business. According to him, India is Google's largest market in Asia. He linked that position to India's economic growth and highlighted Google's long-term investments in local infrastructure. The company's planned $15 billion investment in data centre infrastructure in India forms the foundation of the sovereignty strategy. Kurian suggested that the objective goes beyond serving Indian customers. The infrastructure will also enable India to serve customers in other parts of the world and potentially support broader supply-chain and manufacturing ambitions.
Google is also evaluating opportunities to manufacture components of AI server systems in India. While advanced chips will continue to be produced by global semiconductor partners such as TSMC, Kurian indicated that Google is discussing opportunities around system assembly and related infrastructure with Indian partners.
Another important part of Google's India strategy extends beyond data centres. Kurian revealed that Google has established India as a hub for what it calls "forward deployed engineering." These are specialised engineers who work directly with enterprise customers to help them move AI projects from experimentation into production. Unlike conventional support teams, these engineers function more like product and implementation specialists embedded alongside customers. Their role is to help enterprises build applications on Google's AI stack and deploy them at scale. Google intends for these teams to serve not only India but also customers across Asia. The strategy reflects Google's view that India can become a source of AI expertise rather than merely a consumer of AI technology.
How Google compares with rivals
Google is not the only cloud provider to offer local data residency in India. Microsoft Azure, Amazon Web Services and Oracle Cloud all operate cloud regions in India that allow enterprises to store and process data locally. They also offer AI services through their respective platforms. Microsoft provides AI services through Azure AI and Azure OpenAI. AWS offers foundation models through Amazon Bedrock. Oracle has made generative AI services available through its Indian cloud regions.
The difference lies less in whether AI can be used in India and more in whether the latest frontier models are available for local inference inside India. Model availability often varies by provider, model family and cloud region. New models frequently appear first in the United States before expanding to other geographies.
Google's announcement is significant because the newest Gemini capabilities will now be served from infrastructure located within India itself, with an explicit focus on sovereignty and local processing.
What it means for Indian companies
For Indian enterprises, the biggest benefit is likely to be regulatory comfort. Banks, insurers, healthcare providers, government agencies and large corporations often face internal or regulatory constraints regarding where sensitive information can be processed. Local AI hosting reduces those concerns. It may also improve performance by reducing latency, particularly for applications involving real-time interactions, voice agents and AI-powered customer service.
Overall, the move signals that India is becoming a priority market for advanced AI infrastructure rather than merely a destination for cloud consumption. The combination of local model hosting, large-scale data centre investment, AI engineering talent and potential server manufacturing suggests Google sees India as a strategic AI hub. For Indian companies, that could mean faster access to cutting-edge AI capabilities without having to choose between innovation and data sovereignty.
