Cross-region data replication costs three ways at once: inter-region transfer, duplicated storage, and request charges. The transfer is a one-time charge per byte replicated, typically a cent or two per gigabyte within a provider. The duplicated storage is a recurring monthly charge for every byte kept in the second region, and over time it usually exceeds the transfer cost. Request fees add a smaller third layer. The way to control it is not to stop replicating but to replicate only what truly needs a second region, tier the replica, and align retention to the real recovery objective. Sizing replication to need is part of our 31% average bill reduction.
Last updated: June 2026 · Written by Fredrik Filipsson and reviewed by Morten Andersen · See, Cut, Lock, Run method
Replicating data to a second region is the right call for disaster recovery, low-latency access across geographies, and data-residency requirements. It is also one of the easiest ways to silently double a storage bill, because the cost is split across three meters that rarely get added up together. Teams budget for the transfer and forget the storage recurs, or they replicate everything when only a fraction needs it. This explainer is part of our complete guide to cloud storage and data cost optimization, the cluster pillar it links up to.
What does cross-region replication actually cost?
Cross-region replication costs three things at once: the transfer to move each byte, the storage to keep each byte in the second region, and the requests to perform the replication. The transfer is inter-region data transfer, which on AWS typically runs a cent or two per gigabyte depending on the region pair, per the Amazon S3 pricing page. That charge applies once per byte replicated. The second cost is duplicated storage: every gigabyte you replicate is now stored twice, so the destination region adds a recurring monthly storage charge at its own per-gigabyte rate. The third is requests, the PUT and replication operations that carry the data across, billed per thousand. Of the three, duplicated storage is the one teams underestimate, because it is the only one that keeps charging every month for as long as the replica exists. A one-terabyte dataset replicated continuously pays the transfer once but the second copy of storage forever.
Why is replicated storage the cost that compounds?
Replicated storage compounds because it is a recurring monthly charge, while the transfer is a one-time charge per byte. Think of it in months. In month one, you pay inter-region transfer to seed the replica plus a full month of duplicated storage. In month two and every month after, the transfer cost only applies to new and changed data, but the duplicated storage charge applies to the entire replica again. Over a year, a continuously replicated dataset pays roughly twelve months of second-region storage against a transfer cost that, for stable data, was mostly incurred up front. This is why a replication decision made for a small dataset can look cheap and a replication policy applied across an entire estate can become one of the largest lines on the bill. The lever is selectivity: the cost scales with how much you choose to keep in two places, so replicating only what needs it is the single biggest saving. The transfer side of this same boundary is covered in the sibling guide how to reduce cross-cloud egress costs in a multicloud setup.
Cross-region replication copies current data, including mistakes and deletions, to another region. It protects against a regional outage, not against a bad write or an accidental delete. Versioned backups protect against those. Many teams need both, and paying for both is fine, but treating replication as a backup leaves a recovery gap.
How do I reduce the cost of cross-region replication?
Reduce replication cost by replicating less, tiering the replica, and aligning retention to the real recovery objective. Start with selectivity: classify data by whether it genuinely needs a second region for disaster recovery, latency, or residency, and replicate only that subset rather than defaulting everything to two regions. Next, tier the replica. A disaster-recovery copy that is read only during an incident can sit in a cheaper, colder storage class, so you pay the lower per-gigabyte rate for the duplicate as long as the retrieval latency meets your recovery objective. Compress data before replicating where the format allows, since both the transfer and the duplicated storage scale with bytes. Finally, set the replica retention to the recovery objective rather than keeping every historical version in two places forever. These are the same measure-then-match decisions in our broader storage work, and they route to the same fix in our rightsizing and waste elimination service. The deeper version of the disaster-recovery trade-off appears in our guidance on reducing backup storage costs without losing recovery.
Replicating everything to a second region by default?
Our cost audit classifies what genuinely needs replication, tiers the replicas, and proves the storage and transfer saving against a clean baseline before anything is committed. On the performance model, you pay only from realized savings. No savings, no fee.
Book a cloud cost audit →The Cloud Storage and Egress Cost Playbook includes the replication cost model we use to size cross-region copies to the real recovery objective.
Frequently asked questions
What does cross-region replication actually cost?
Cross-region replication costs three ways at once: inter-region data transfer for every byte replicated, duplicated storage in the destination region for every byte kept, and request charges for the replication operations. The storage cost recurs every month, so over time the duplicated storage usually exceeds the one-time transfer cost.
Is replicating to another region the same as a backup?
No. Replication copies current data to another region for availability or residency, but it also copies mistakes and deletions, so it is not a substitute for versioned backups. Many teams pay for both replication and backup because each serves a different recovery need.
How do I lower cross-region replication cost?
Replicate only the data that genuinely needs a second region, tier the replica to a cheaper storage class where retrieval latency is acceptable, compress before replicating, and align the replica retention to the real recovery objective rather than keeping everything forever in two places.
Is inter-region transfer cheaper than internet egress?
Yes, usually. Inter-region transfer within a provider typically costs a cent or two per gigabyte, less than the roughly nine cents per gigabyte for internet egress on AWS, but it is still a real recurring charge on continuous replication and it stacks on top of the duplicated storage.
The short version
Cross-region replication costs three ways: inter-region transfer to move the data, duplicated storage to keep it, and request fees for the operations. The duplicated storage recurs monthly and is the cost that compounds, so it usually dwarfs the transfer over time. Control it by replicating only what needs a second region, tiering the replica, and aligning retention to the recovery objective. When you want replication sized to real need and the saving proven down, that is the work our rightsizing and waste elimination service delivers.
Cloud pricing and service behavior change frequently. Verify the specifics in this guide against the providers’ own current documentation and the FinOps Foundation: FinOps Foundation Framework ↗ and FinOps Rate Optimization capability ↗. This article also reflects Cloud Cost Room’s hands-on, vendor-neutral engagement experience.