These are jurisdictional and regional claims, not country-level hosting assertions. A geographic endpoint can route across multiple facilities within its documented boundary.
Each row states exactly what the provider documents. API access, request storage, inference processing, and jurisdictional routing are separate claims.
Model Studio Frankfurt EU deployment scope restricts inference to the EU; model availability remains region-specific.
AWS geographic inference profiles can keep supported model inference within the EU; model-specific destination regions vary.
Google documents EU multi-region endpoints for supported Google and partner models.
Official evidence: Google Vertex AI model endpoint locations
Microsoft Foundry documents project regions in France, Germany, Italy, Spain, Sweden, and other EU locations; model availability varies.
Mistral EU regional endpoint processes supported model inference in EU and EFTA data centers; control-plane data is separate.