What is Actually Changing?
OpenAI has expanded its services in India by launching local API processing. Previously, the company offered data residency, which meant eligible customers could store their data within India. The new announcement on November 12, 2026, takes this a significant
step further. Now, for supported workloads, not only is the data stored in India, but the entire computational process—known as inference—also happens on servers located within the country. This means that when an eligible Indian business sends a request to the API, the query is processed and the AI-generated response is created without the data leaving India's borders. At launch, this capability is available for the GPT-6 Luna model through the Responses and Chat Completions APIs.
Why Local Processing is a Big Deal
For many Indian companies, especially in sensitive sectors like financial services and healthcare, data sovereignty is not just a preference but a strict regulatory requirement. India's Digital Personal Data Protection (DPDP) Act, along with specific mandates from bodies like the Reserve Bank of India (RBI), governs how and where user data can be handled. Until now, the need for API requests to be processed on international servers was a major barrier to the adoption of powerful AI tools for core business functions. According to Nitin Bawankule, OpenAI's Head of Enterprise in India, this prevented some businesses from deploying AI for important work. By ensuring the entire process stays in-country, OpenAI addresses these data localisation concerns, giving businesses the control and compliance assurance they need.
The Benefits for Indian Businesses
The most immediate benefit is regulatory compliance, which unlocks AI for previously blocked use cases like analysing sensitive customer information or drafting internal documents in regulated industries. Beyond compliance, local data processing can lead to significant performance improvements. Reducing the physical distance data has to travel—from India to servers in Europe or the US and back—can lower latency. This results in faster response times from the AI, making applications feel more fluid and responsive, which is critical for real-time uses. This move also builds trust with Indian users who are increasingly conscious of their data privacy. Furthermore, having a local data processing option can give companies a competitive edge when bidding for government contracts, which often mandate in-country data handling.
Who Can Use It and How?
This new capability is available for eligible businesses and developers using the OpenAI API on a pay-as-you-go basis. To get started, customers need to create a new API project specifically configured for India. In the project settings, they must select India for both data storage and processing. Once set up, they will use a dedicated India regional endpoint (in.api.openai.com) to ensure their requests are routed to and handled by the local infrastructure. Initially, the local processing is only available for the GPT-6 Luna model and applies to direct API usage for text and, with approval, image inputs. It's important to note that this doesn't automatically apply to all of OpenAI's products, like ChatGPT, but is a specific feature for API customers who opt-in.
The Broader Context of 'OpenAI for India'
This launch is part of a wider strategy called 'OpenAI for India', which was announced earlier in the year to accelerate AI adoption across the country through infrastructure, enterprise solutions, and skills development. A key component of this strategy was OpenAI's partnership with the Tata Group to develop domestic AI data centre capacity, which laid the groundwork for this local processing launch. India is not alone in getting this feature; OpenAI has been rolling out similar regional processing capabilities in other markets, including Australia and Japan, signalling a global trend towards accommodating local data laws. This also follows a similar move where OpenAI models were made available for in-country inference through Amazon Web Services' Bedrock platform.
















