OpenAI brings local data residency to Asia

For enterprise IT leaders and university admins, adopting frontier AI usually dies at the legal desk with one simple question: where exactly is our data being stored?

OpenAI is tackling that compliance hurdle head-on in the Asia-Pacific region. The company announced that customers in Japan, India, Singapore, and South Korea can now store their at-rest data locally within their respective borders.

What gets stored locally, and how to set it up

The local storage guarantee applies to ChatGPT Enterprise, ChatGPT Edu, and the API Platform.

For ChatGPT Enterprise and Edu, administrators setting up new workspaces can select their home country as the primary storage location. That covers conversations with ChatGPT, custom GPT configurations, user prompts, uploaded documents, and files across text and vision modalities.

Developers using the API get the same flexibility. Eligible accounts can simply specify their region when creating a new Project inside the API dashboard, pinning all associated at-rest data to local servers.

OpenAI paired the rollout with a reminder of its enterprise privacy stack. Data at rest is secured with AES-256 encryption, while data in transit uses TLS 1.2 or higher. The company reiterated that it never trains models on enterprise or API customer data by default, and that its systems comply with SOC 2 Type 2, CSA STAR, and standard Data Processing Addenda.

The feature formalizes protections for hundreds of organizations already running on OpenAI across Asia, including SoftBank, Kakao, Grab, and Singapore Airlines. A later update in late 2025 noted that global at-rest data residency options have since expanded to additional regions worldwide.

Why it matters

This solves one of the biggest procurement roadblocks for regulated sectors across Asia. Banks, healthcare networks, and public universities often face rigid local data sovereignty rules that previously ruled out direct contracts with OpenAI, pushing many buyers to third-party cloud brokers or Microsoft Azure wrappers. Offering native data residency lets OpenAI close enterprise deals directly.

It also squeezes regional AI challengers. In markets like Japan and South Korea, domestic foundation model startups frequently pitch local data hosting as their primary moat against Silicon Valley giants. When OpenAI meets those exact compliance specs, local alternatives have to compete purely on price and language nuance.

That said, the fine print is worth noticing. OpenAI specifically specifies data residency at rest. That means your prompts and files sit on local disks when idle, but the actual GPU inference processing may still route through global infrastructure. Existing enterprise accounts will also need to spin up fresh workspaces or projects to take advantage of the switch, which means data migration will still require some manual lift.