ChatGPT maker OpenAI on Thursday announced that data for its application programming interfaces (APIs) in India would now be processed within the country. Inference for these supported APIs would also be done within India, the company said.
“Available initially for the GPT-6 Luna model, this offering builds on OpenAI’s existing data residency options in India and helps organisations adopt AI for work that requires in-country processing,” OpenAI said.
The move is likely to be beneficial for companies in the financial technology sector, where the sectoral regulator, the Reserve Bank of India, has mandated local processing and storage of data.
“Local inference gives them more control over where supported workloads run and helps address a practical barrier to adoption. It gives more businesses a way to put OpenAI’s tools to work, and gives Indian developers and startups more scope to build for their own, and enterprise needs,” said Nitin Bawankule, head of enterprise in India.
In May last year, OpenAI started offering data residency in India to Indian ChatGPT Enterprise, ChatGPT Edu and OpenAI API Platform users to help companies using these products meet local data sovereignty requirements when using the company’s products in their businesses and building new solutions with AI.
The Friday announcement extends beyond that, as the processing of an API request and the model computation that generates its response now take place on servers within the country.
“Supported customer content is also stored in-country, helping businesses bring AI to text-based work such as summarising documents, drafting responses, and helping employees analyse information,” the company said.
In addition to starting a data residency programme in India last year, the world’s leading frontier AI model development company also introduced OpenAI for India, a new initiative for AI infrastructure, enterprise adoption, skills and local partnerships.
As part of that announcement, OpenAI also tied up with the Tata Group to develop domestic AI data centre capacity.
