Bengaluru: Google is expanding its sovereign AI offerings in India, introducing in-country AI processing and fully air-gapped deployments for enterprises as demand grows for secure, locally hosted artificial intelligence.Speaking at Google I/O Connect India 2026 held in Bengaluru on Tuesday, Richard Seroter, chief evangelist at Google Cloud, said Google is enabling in-country processing and inference for Gemini 3.5 Flash, ensuring enterprise AI workloads and data remain within India. The company is also bringing Gemini on Google Distributed Cloud to the country, allowing organisations to deploy frontier AI models entirely within their own infrastructure, disconnected from the public internet.The offering is aimed at highly regulated sectors such as defence, government and critical infrastructure that require the highest levels of security and compliance.Seroter, however, said sovereign AI is not a one-size-fits-all proposition. While fully isolated deployments are essential for a small set of customers, he said most enterprises would continue to favour public cloud with in-country data controls.“If you can use the public cloud while keeping your data in-country, you get the benefits of scale without having to buy and manage your own infrastructure,” he said. “For most customers, that’s still the preferred option.”Dedicated, air-gapped deployments, he added, are typically a last resort for organisations that cannot connect to the internet because of stringent regulatory or national security requirements.“We have a pretty unique sovereign story at Google,” Seroter said. “Whether customers need local processing, sovereign controls or completely disconnected infrastructure, we want to give them the flexibility to choose the level of sovereignty that fits their business.”Beyond infrastructure, Seroter said India has become central to Google’s global AI strategy.“I don’t think of India as an emerging market anymore. It has already emerged,” he said, noting that India is now among Google’s top three developer ecosystems worldwide.He also argued that AI is transforming, rather than replacing, software engineering.“Developers are becoming orchestrators,” Seroter said. “AI makes execution easier, but human judgement becomes even more important. People will spend more time defining problems, evaluating outputs and applying domain expertise.”While concerns persist over organisations becoming overly dependent on AI, Seroter stressed that accountability remains with humans.“AI can assist, but it doesn’t assume responsibility,” he said. “Whether it’s software, financial advice or legal work, the human remains accountable for the final outcome.”
