From Local Storage to Local Processing
OpenAI's presence in India has taken a significant leap forward. While the company has offered local data storage in India since May 2025, its latest announcement on November 12, 2026, adds a crucial new layer: local processing. Previously, an Indian
company's data could be stored at rest within the country, but the actual computational work—known as inference—might have happened elsewhere. Now, for eligible customers and supported models, both the API request and the AI model's computation to generate a response can happen on servers based in India. This end-to-end, in-country data lifecycle is a critical development for companies navigating India's data landscape.
Why In-Country Processing Is a Big Deal
The primary driver for this move is data localisation and compliance. Many Indian businesses, especially in highly regulated sectors like financial services and fintech, have been cautious about adopting AI for critical operations. The need to have sensitive information processed within national borders, as mandated by regulators like the Reserve Bank of India (RBI), was a major barrier. As OpenAI's Head of Enterprise in India, Nitin Bawankule, noted, some businesses had not fully deployed AI because of this specific requirement. Local inference gives these organisations greater control and helps them meet strict data sovereignty rules, unlocking a new range of applications like summarising sensitive documents or analysing internal data.
Who Can Use It and For Which Models?
This feature is not a default setting for all users; it is an enterprise-focused capability. To be eligible for regional processing in India, organisations must have active Zero Data Retention (ZDR) controls. At launch, the local processing capability supports the GPT-6 Luna model via the Responses and Chat Completions APIs. This covers applications involving text and image inputs with text outputs. Businesses that need this feature must create a new API project, specifically select India for both storage and processing, and use the designated regional endpoint: in.api.openai.com.
The Broader Strategy: 'OpenAI for India'
This launch is part of a wider strategic initiative called 'OpenAI for India', announced in February 2026. The program focuses on building local AI infrastructure, accelerating enterprise adoption, and developing skills. A cornerstone of this strategy is a partnership with the Tata Group to develop domestic AI-ready data centre capacity, ensuring long-term sovereign AI capabilities. This collaboration aims to deliver lower latency and enhanced security for mission-critical workloads. The move also aligns India with other key markets like Japan and Australia, which recently received similar local processing capabilities from OpenAI.
What About the Cost and Other Details?
Using regional endpoints can come with additional costs. According to OpenAI's pricing guidelines, data residency endpoints for eligible models released after March 5, 2026, may have a 10 percent price uplift. It's also important to note what isn't covered. Data residency and processing apply only to specific 'customer content' sent through the API. It does not cover other data like account details, billing information, or usage statistics, which may still be processed and stored outside India. Businesses considering this option should carefully review OpenAI's data control documentation and consult with their sales team to confirm eligibility and understand the full implications for their specific use case.
















