Home/Library/Reduce Cross-Cloud Egress
How-to · Data, Storage & Egress Cost · Updated June 2026

How to Reduce Cross-Cloud Egress Costs in a Multicloud Setup

Multicloud sounds like flexibility until the data transfer bill arrives. Internet egress runs about $0.09 per gigabyte on AWS for the first tier, and every byte crossing a cloud boundary pays it. Co-locate chatty workloads, cache repeated reads, and route steady flows over private links to stop paying twice for the same data.

Key takeaways

Cross-cloud egress is an architecture cost, not a line item you can simply switch off. Providers charge to send data out, around $0.09 per gigabyte on AWS for the first tier, while inbound is free, so multicloud traffic pays every time it crosses a boundary. Cut it by mapping where egress actually flows, co-locating chatty services so their traffic stays internal, caching and compressing at the boundary, and routing steady high-volume flows over private connectivity that carries lower per-gigabyte rates. The largest savings come from not sending the data at all. In the environments we optimize, egress is one of the most overlooked lines, and addressing it is part of our 31% average reduction.

Last updated: June 2026 · Written by Fredrik Filipsson and reviewed by Morten Andersen · See, Cut, Lock, Run method

Egress is the cloud charge teams understand least and pay most for by accident. Every major provider lets data in for free and charges to send it out, so a multicloud architecture that moves data between providers, between regions, or out to users is paying a meter that compounds with every chatty integration. The fix is rarely a single setting; it is a handful of architectural choices that keep data from crossing the expensive boundaries in the first place. This guide is part of our complete guide to cloud storage and data cost optimization, the cluster pillar it links up to.

Why is cross-cloud egress so expensive?

Cross-cloud egress is expensive because providers price outbound data transfer to discourage moving data off their platform, while charging nothing to bring it in. On AWS, data transfer out to the internet is free for the first 100 gigabytes a month and then about $0.09 per gigabyte for the first tier, declining at higher volumes, per the AWS data transfer pricing. Inter-region transfer within AWS is cheaper, typically a cent or two per gigabyte, and inter-Availability-Zone traffic is about $0.01 per gigabyte each way. The asymmetry is the trap: a multicloud design that reads data from one provider and processes it in another pays the outbound meter on every read, and because inbound is free, the cost is easy to miss when you only look at the receiving side. The expensive boundaries, in order, are internet and cross-cloud egress, then cross-region, then cross-AZ. Architect to keep traffic on the cheap side of each boundary.

How do I reduce cross-cloud egress step by step?

Reduce egress by keeping data from crossing the expensive boundaries, then making the crossings that remain smaller and private. Work through these steps:

  1. Map where the egress actually flows. Use each provider's cost and usage report filtered to data transfer usage types to find which services and regions send the most data across cloud and region boundaries. A few workloads usually drive most of the bill.
  2. Co-locate chatty workloads. Move services that talk constantly into the same cloud and region so their traffic stays internal and cheap. A database and the app that hammers it should never sit on opposite sides of an egress boundary.
  3. Cache and compress at the boundary. Put a CDN or cache in front of repeated reads, and compress payloads, so the same data crosses the boundary fewer times and in fewer bytes. Repeated reads of static data are pure waste.
  4. Use private connectivity for steady flows. Route high, steady cross-cloud traffic over private interconnects or peering, which carry lower per-gigabyte rates than internet egress once volume covers the fixed port cost.
  5. Set egress budgets and alerts. Add anomaly alerts on data transfer usage types so a new chatty integration is caught in days, not at month end when the invoice lands.

This is the same waste-elimination work in our rightsizing and waste elimination service. The replication side of cross-boundary cost, where copies of data multiply the transfer bill, is covered in the sibling guide what is the real cost of data replication across regions.

The cheapest byte is the one you never send

Most egress savings come from architecture, not discounts. Co-locating a service with the data it reads, or caching a hot dataset at the edge, removes the transfer entirely. Private links and committed rates help with the traffic that genuinely must cross, but they are the second lever, not the first.

Does a CDN or private connectivity reduce egress?

Both reduce egress, but for different traffic: a CDN cuts repeated reads, and private connectivity cuts the rate on steady high-volume flows. A content delivery network caches frequently requested objects close to users, so repeat requests serve from the edge instead of crossing the cloud boundary again, which makes it highly effective for static assets and popular content but useless for one-time bulk transfers. Private connectivity, such as a dedicated interconnect or cloud peering, carries a lower per-gigabyte rate than standard internet egress and is the right tool when large, continuous volumes flow between two clouds or between a cloud and a data center. The decision is about traffic shape: cache what is read repeatedly, and put a private link under what flows steadily and large. For the detailed version of the private-link decision, see our work on cutting egress with private connectivity and peering.

Is multicloud quietly taxing you on every byte?

Our cost audit maps egress by service and boundary, co-locates the chatty workloads, and proves the 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 →
Go deeper · free playbook

The Cloud Storage and Egress Cost Playbook includes the egress mapping and reduction checklist we run, so multicloud traffic stops draining the bill.

Frequently asked questions

Why is cross-cloud egress so expensive?

Cross-cloud egress is expensive because providers charge to send data out to the internet, typically around $0.09 per gigabyte on AWS for the first tier, while inbound is usually free. Every byte that leaves one cloud for another crosses that meter, so chatty multicloud architectures pay repeatedly for the same data.

How do I find what is driving my egress bill?

Use each provider's cost and usage report filtered to data transfer usage types. They show which services and regions generate the most outbound traffic, so you can target the few workloads responsible for most of the bill rather than guessing across the whole estate.

Does a CDN reduce cross-cloud egress?

Yes for repeated reads. A CDN caches frequently requested data at the edge, so repeat requests serve from the cache instead of crossing the cloud boundary again. It cuts egress on static and popular content, though it does not help one-time bulk transfers.

Is private connectivity cheaper than internet egress?

For steady, high-volume flows, usually yes. Private interconnects and peering carry lower per-gigabyte rates than standard internet egress, so they pay off once cross-cloud traffic is large and continuous enough to cover the fixed port cost. For low or bursty traffic, internet egress can still be cheaper.

The short version

Cross-cloud egress is an architecture cost. Providers charge to send data out, around $0.09 per gigabyte on AWS for the first tier, so multicloud traffic pays every time it crosses a boundary. Map where egress flows, co-locate chatty services, cache and compress at the boundary, and put private links under steady high-volume flows. The cheapest byte is the one you never send. When you want egress mapped and the saving proven down, that is the work our rightsizing and waste elimination service delivers.

Primary sources & further reading

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.

Co-founder of Cloud Cost Room and a FinOps Certified Practitioner, with 20 years in IT and cloud cost optimization across AWS, Azure, Google Cloud and OCI. More about Fredrik →

More from the Data, Storage & Egress Cost cluster

See every guide in the Data, Storage & Egress Cost cluster →

The Cloud Cost Brief

Cloud pricing moves. We tell you when it matters.

New commitment instruments, FOCUS changes, hyperscaler pricing shifts, and the plays that actually move a bill. No schedule, no filler.

Subscribe · Work email only