AWS MCP Server adds six hosting Regions

AWS announced on October 2, 2026 that its managed AWS MCP Server is available in six additional Regions: Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Ireland), Europe (London), and US West (Oregon).

Part of the Agent Toolkit for AWS, the service is a managed Model Context Protocol server that gives AI coding agents a single interface to AWS APIs. AWS says agents can use it to discover and call services without building and maintaining separate integrations for each service, according to the announcement.

Eight hosting Regions, broader service access

The six additions bring the announced hosting footprint to eight Regions, alongside US East (N. Virginia) and Europe (Frankfurt).

That footprint is distinct from the server’s service reach. The availability notice explicitly states that AWS MCP Server can access services in all commercial AWS Regions, while the server itself runs in the eight listed locations.

For deployment reviews, this creates two separate questions: where does the coding agent connect to the MCP server, and which Regions will its requested actions affect? Choosing a nearby endpoint should not be treated as a guarantee that every resource involved in a workflow is in that same Region.

Latency and residency need separate checks

AWS says the expansion lets customers run the server closer to their developers with lower latency and keep request data within the Region to meet data residency requirements. Those are AWS’s stated benefits; the announcement does not quantify the latency improvement.

Teams should measure latency against their own agent workflows and assess residency requirements across the complete request path and target services. The expansion provides more endpoint locations, but the endpoint’s location alone is not a complete assessment of where a workflow’s data and resources reside.