AWS MCP Server expands to six new regions, bringing AI agent tooling closer to developers

AWS MCP Server lands in six more regions

The AWS MCP Server, a managed service that lets AI agents work with AWS through the Model Context Protocol, is now available in six additional regions: Singapore, Sydney, Tokyo, Ireland, London and Oregon. According to a dev.to post by Aaron Hunter, the expansion brings the total footprint to eight regions, joining the two original locations of N. Virginia and Frankfurt.

What the service actually does

For those who have not encountered it before, the MCP Server acts as a remote, managed intermediary between AI coding assistants and an AWS account. Hunter names Claude, Codex, Cursor and Kiro as typical clients. Through the server, agents can run AWS CLI commands, search AWS documentation and apply curated skills maintained by AWS for common tasks.

The server is one piece of a larger package called the Agent Toolkit for AWS, which AWS assembled for teams building agents that operate against real cloud infrastructure rather than only generating code in the abstract.

Why regional coverage matters

Two practical consequences follow from the expansion.

The first is latency. Teams can now run the server and the Agent Toolkit physically closer to the infrastructure they are building against. Hunter's illustration is a London-based team pointing its agents at a local endpoint to build features, inspect running workloads and debug failures, without routing every request halfway around the world.

The second is data residency. Because request data stays within the chosen region, organisations whose compliance obligations require processing to remain in a specific jurisdiction gain more deployment options than the original two-region lineup allowed.

One caveat from the post is worth flagging: regional placement does not restrict what an agent can reach. Even when the server runs in, say, Tokyo, it can still access services across all commercial AWS Regions. The chosen region determines where the server executes and where request data resides, not which resources an agent may touch.

The full region lineup

With this round of additions, the supported regions are:

• N. Virginia
• Frankfurt
• Oregon
• Singapore
• Sydney
• Tokyo
• Ireland
• London

How to get started

For teams not yet using the MCP Server or the wider Agent Toolkit, the post recommends enabling them if you are building on AWS. The entry point is a single command, available from AWS CLI version 2.35 onward:

aws configure agent-toolkit

Why it matters

AI coding assistants are increasingly expected to do more than write code: they provision resources, query running systems and help debug live failures. MCP has emerged as the standard plumbing for that kind of access, and AWS offering a managed server means teams do not have to host and secure their own gateway to the cloud.

Until now, that convenience came with a geographic compromise. Two regions, both in the United States and Europe, meant Asian-Pacific teams absorbed cross-continent latency and teams with strict data residency rules had limited options. Eight regions spanning Asia-Pacific, North America and Europe turns the server from an early-adopter convenience into something globally practical for everyday development workflows.

The move also signals how AWS views agent-facing infrastructure: as a first-class regional service alongside compute and storage, rolled out with the same coverage playbook as any other managed offering. For developers wiring assistants like Claude, Cursor or Kiro into AWS-heavy projects, the setup decision just got simpler — pick the region you already build in.

One thing the source does not cover is pricing changes or capacity differences between regions, so teams with heavy agent workloads may want to verify behaviour in their specific region before standardising on it.