OpenAI adds UK data residency for API customers
A new gb.api.openai.com endpoint keeps eligible customer content stored in the UK, joining nine other regional endpoints OpenAI already offers.
OpenAI released UK data residency for the API on 24 October, adding a gb.api.openai.com endpoint to the nine it already runs for the US, EU and EEA, Australia, Canada, Japan, India, Singapore, South Korea and the UAE. Route a project’s traffic to it and, per OpenAI’s own wording, “your customer content… for that project will be stored at rest in the selected region to the extent the endpoint requires data persistence.”
The endpoint covers most of the API surface: chat completions, Responses, Decisions, embeddings, batches, audio, image generation and editing, fine-tuning, and file operations. Account, billing and analytics data can still process outside the region. Access is not self-serve: OpenAI says to “contact our sales team to see if you’re eligible,” and non-US regions need sign-off on abuse monitoring controls plus a Modified Retention amendment. There’s also a cost: a 10% uplift on data residency endpoints for models released on or after 5 March 2026.
Why it matters: if you’re building for UK customers under data protection rules that expect content to stay in-country, this is a concrete lever rather than a promise buried in a DPA, you point traffic at a different subdomain. The approval process and the 10% surcharge mean it’s worth checking whether you actually need regional storage before switching, rather than defaulting every project to it.