AWS Cross-Region Data Transfer Cost Calculator

Cross-region transfer is usually directional, not symmetric

This page is for one of the easiest cloud charges to misread: traffic that moves between regions. The billing event is often directional, so a clean estimate starts with region pair and traffic direction, not just total bytes moved.

• Steady-state replication: predictable monthly copy volume from the primary path.
• Backfill or migration: one-time bulk transfer that should never be mixed into the normal month.
• DR or failover: rare but expensive spikes that need their own peak scenario.

What people usually miss on cross-region bills

The transfer line itself is only one part of a multi-region decision. Teams often estimate the copied GB but miss the fact that replica storage, destination reads, and backlog catch-up events all have different billing behavior. That is why this page should stay focused on the transfer line and hand off the adjacent lines to other tools.

• Directionality: confirm whether the cost is charged on the outbound side only.
• Event separation: migrations and backlog replays belong in a separate peak case.
• Neighboring charges: replica storage and destination-side processing should not be hidden in this number.
• Compression changes: payload optimization may reduce copied GB without changing request count.

What to capture before you trust the estimate

• Identify the specific source and destination region pair instead of using a generic cross-region assumption.
• Split steady replication from migrations, backfills, DR drills, and failover recovery windows.
• Keep cross-region traffic separate from internet egress and cross-AZ transfer even if the same system generates all three.
• Review whether the real business question is transfer-only or a wider multi-region cost model.

Steady-state vs event-driven region transfer

How to review the first real cross-region bill

• Check whether the billed direction matches the direction you modeled for the region pair.
• Keep one-time migrations and failover windows outside the recurring baseline after the event is complete.
• Investigate destination storage, replay traffic, or read behavior before blaming the transfer rate itself.

Next steps