समाचार Samāchār · Tech & AI · 12 November 2026

OpenAI now runs its models inside India. Local inference is not yet sovereign AI

OpenAI launched API inference residency in India on 12 November. Requests to GPT-6 Luna can now be processed and stored in the country. The scope is narrow at launch.

An Indian company can now send a request to OpenAI’s API and have the model run in India. OpenAI launched API inference residency in India on 12 November. Until now OpenAI could store an Indian customer’s data in the country. The computation could happen elsewhere. Now both can stay at home.

The scope is narrow at launch. It covers one model, GPT-6 Luna, through the Responses API and the Chat Completions API. Customers create a new API project, choose India for both storage and processing, and use the India regional endpoint. India joins Australia and Japan as markets with this option. OpenAI has named the Tata Group as its partner for domestic AI data centre capacity.

Nitin Bawankule, OpenAI’s Head of Enterprise for India, named the barrier plainly. Some Indian businesses have not deployed AI for important work because they need the information processed within the country, he said. Analytics India Magazine reported the move could help financial services firms and other regulated businesses meet data localisation requirements. Banks and insurers are the obvious first users.

This is a विजय Vijay for the demand side of Indian AI. The market is large. Inc42 puts weekly ChatGPT users in India above 10 crore. A market that size pulls compute into the country. But inference residency is not sovereignty. The model, its weights and its price are still decided elsewhere. The racks are capacity in India. The intelligence is not yet Indian.

The lesson for builders runs two ways. Local processing removes an excuse for regulated sectors to keep waiting. And the same demand is the market for India’s own foundation models, its own compute and the power to run them. A country that hosts others’ inference should also run its own.

What a founder takes from it: For founders selling AI to banks, insurers and hospitals, where the data is processed is no longer a reason to wait. The larger prize is the same demand served by Indian models on Indian compute and Indian power.