Backfrom Anthropic’s own policies · verified 2026-09-02
Anthropic: data residency & retention
Headquarters
🇺🇸 United States (US)
Where inference runs
US
The inference_geo API parameter (Claude 4.6+ models) pins inference to US infrastructure at a 1.1x price; the default 'global' setting may run inference in any available geography. Data at rest is US-only.
Region pinning
Yes, self-serve
Self-serve per-request and per-workspace inference-geo controls exist, but only 'us' and 'global' are offered — there is no EU pinning option.
Prompt retention
Retained · 30 days
API inputs and outputs are automatically deleted within 30 days; the newest frontier models require 30-day retention, and content flagged by trust-and-safety systems can be kept up to 2 years.
Trains on API data
No
Retained API data is never used for model training without the customer's express permission, per Anthropic's Commercial Terms.
Zero data retention
Available
ZDR must be requested through sales and enabled per organization; the newest frontier models are excluded from ZDR unless expressly authorized.
Compiled from Anthropic’s published privacy policy, terms, and documentation on 2026-09-02. “Not disclosed” means Anthropic publishes nothing on the point. Policies change without notice, so verify against the sources above before relying on this for compliance decisions.