OpenAI announced on Thursday that businesses and developers in India can now have their API requests processed inside the country, with the model itself running on local infrastructure. The capability launches first for GPT-6 Luna, through the Responses and Chat Completions APIs, and extends OpenAI’s existing Indian data residency from storing data locally to actually computing on it there. It is aimed squarely at organisations that have held back from deploying AI for important work because the information involved must stay within India’s borders.
What actually changed
Until now, OpenAI’s Indian data residency covered storage. Customers could keep their data in the country, a setup introduced in May 2025. The new step covers the other half of the journey: both the processing of an API request and the model computation that generates the response take place in India for supported workloads, with customer content stored there too.
Getting started is a configuration choice rather than a new product. Customers create a new API project, select India for both storage and processing, and use the India regional endpoint. For now the service covers API usage only, and only GPT-6 Luna inputs and outputs are supported, so teams running other models will wait for OpenAI to widen the list.
Why India, and why now
India is the market where this matters most. The country recently passed the United States as ChatGPT’s biggest market, and OpenAI has spent 2026 deepening its footprint there, from an OpenAI for India initiative spanning infrastructure, enterprise adoption and skills to a partnership with Tata Group to build domestic AI data-centre capacity.
Regulation is the other driver. Financial services firms face data localisation rules from the Reserve Bank of India, and OpenAI’s India enterprise chief, Nitin Bawankule, said some businesses had not fully deployed AI for important work because the information involved must be processed within the country. Local inference, he said, gives them more control over where supported workloads run.

Our Read
Residency is becoming the enterprise moat in AI. Google and Anthropic face the same localisation pressure in every regulated market, and the vendor that can credibly promise data never leaves the country wins the deals that matter — banks, insurers, hospitals and governments. OpenAI’s move is less about one model than about removing the last objection in its largest market.
The launch is narrow, and that is worth saying plainly. One model, direct API use only, and no disclosed pricing or availability numbers — the announcement says eligible customers, not every customer. The test will be how quickly OpenAI adds models and whether the India endpoint performs as well as its global ones.
Still, the direction is clear. OpenAI’s enterprise push produced its first $10 billion quarter, and that growth now runs through local infrastructure in the markets that demand it — a pattern rivals will have to match in India, Australia and Japan, the latest markets to get the capability.
Why It Matters
For Indian businesses, the practical barrier is gone: regulated work can now run on OpenAI’s API without data leaving the country. For OpenAI, it is a foothold in its most important growth market, and a template it can repeat wherever regulators draw the same line.
Background
OpenAI began offering data residency in India in May 2025, covering storage for ChatGPT Enterprise, ChatGPT Edu and API Platform customers. In August, its models also became available for in-country inference through Amazon Web Services’ Bedrock platform. India, Australia and Japan are the latest markets to receive OpenAI’s API inference residency.
Source: The Hindu’s report on the announcement and The Economic Times coverage


0 Comments