OpenAI launches India API inference residency for enterprises

By Grace Kim

OpenAI now offers India-based API inference for GPT-6 Luna in a market where, according to the company, more than 100 million people interact with ChatGPT weekly. Initial availability covers GPT-6 Luna. For supported workloads, API requests and the computing required to generate responses can take place within India. The service extends OpenAI’s existing data-residency options in the country.

What changes for Indian enterprises?

Customers must take three steps to use the offering:

• Create a new API project.
• Select India as the location for both data storage and processing.
• Use the corresponding India regional endpoint.

OpenAI India enterprise head Nitin Bawankule said local inference gives businesses greater control over where their data is processed. It could also move more organizations from AI experimentation into daily operations.

“Some Indian businesses have not fully deployed AI for important work because they need the information involved to be processed within the country,” Bawankule said.

Why does data localization matter?\n

Global models can offer advanced capabilities, but enterprises handling sensitive information need to know where their data goes. For banks, financial institutions and government agencies, data storage and processing before deployment require strict security controls.

Residency and infrastructure requirements can limit international frontier AI models in production environments. OpenAI’s India offering addresses that constraint for supported GPT-6 Luna workloads. Initial availability does not extend to every OpenAI model.

India now joins Australia and Japan among the markets where OpenAI has introduced API inference residency. The expansion gives customers another way to meet domestic processing requirements without moving sensitive workloads across borders.

How does competition in India change?

In August, Anthropic announced plans to bring in-country inference for its Claude AI models to India through Amazon Bedrock, AWS’s managed service for building generative AI applications. Anthropic has since made similar data-residency capabilities available for certain models on AWS Bedrock.

The move targets residency and infrastructure requirements that have kept banks, financial institutions and government agencies from deploying international AI models in production. For Indian startup Sarvam AI, locally hosted inference has served as an important differentiator against global model providers.

OpenAI and Anthropic now offer similar capabilities in India, which could narrow Sarvam’s advantage. Their entry also gives local enterprises a broader choice of frontier-model providers that can process requests locally.

Can OpenAI monetize its weekly user base?\n

OpenAI describes India as its largest market and says more than 10 crore people interact with ChatGPT weekly. That is more than 100 million weekly users, giving the company a large audience for paid API and business products. API inference residency adds a route to convert enterprise demand.

No local inference price, India API revenue or new capacity target accompanied the announcement. The commercial test is therefore conversion of that user base into production deployments, rather than reach alone.

What does the new ChatGPT business layer add?

OpenAI has introduced Businesses in ChatGPT, a beta offering that lets companies claim and manage profiles through a new Business Hub. Businesses can update their information, add product feeds and access aggregated insights into how users discover and engage with them on ChatGPT.

Those insights cover business mentions, website clicks and discovery topics across conversations that cite company websites. Businesses in markets where ChatGPT is available, including India, can claim profiles through Business Hub, subject to eligibility and verification requirements.

OpenAI has also introduced Business Agents, which allow users to interact with businesses, explore their offerings and ask product-related questions directly inside ChatGPT. Only eligible ecommerce and online-service businesses in the United States can use the feature initially.

What infrastructure supports the India push?

Earlier this year, OpenAI partnered with Tata Group to develop domestic AI data-center capacity. OpenAI and Tata Group plan to develop that capacity; the announcement did not specify an investment value, server count or capacity target. OpenAI has also extended data-residency capabilities for certain models through AWS Bedrock.

Its broader India strategy combines investments and partnerships for domestic AI infrastructure with localized API execution and business-facing tools. The commercial effect depends on how quickly the Tata partnership adds usable capacity and whether residency wins sensitive workloads from Anthropic and Sarvam AI.

The next test is whether OpenAI can add usable domestic capacity and convert weekly reach into paid deployments as Anthropic and Sarvam AI narrow the remaining localization advantage.