Cloud Egress Fees Explained: How Data Transfer Costs Change by Provider

Cloud egress fees are one of the easiest ways for infrastructure costs to creep up quietly. Uploading data into a cloud platform is often free, but moving that same data out to the internet, to another region, or back to your own data center can trigger outbound transfer charges. For teams planning launches, migrations, or high-traffic applications, understanding those charges is just as important as choosing compute or storage.

This reference breaks down how cloud egress fees work, where bill shock usually appears, and how pricing patterns differ across AWS, Azure, Oracle, and Backblaze. It is designed to be updated as vendor pricing changes, so treat the figures below as a practical snapshot rather than a permanent rate card.

What cloud egress fees are and when they apply

Egress is outbound data transfer. In cloud billing, it usually means data leaving a provider’s environment, whether that destination is the public internet, another cloud region, or sometimes an external data center. Ingress, by contrast, is data coming into the cloud. Based on the supplied evidence, ingress is generally not charged in the same way as outbound transfer.

The important thing for buyers is that “egress” is not one single line item. A transfer from storage to a CDN, from one region to another, or from cloud services back to an on-premises system may be billed differently. That is why a cloud bandwidth comparison needs to look at the transfer path, not just the headline price per GB.

Quick comparison of outbound transfer pricing by provider

How providers define egress differently

• Region-to-region transfers may cost differently from region-to-internet transfers.
• Some services publish simple rates, while others rely on calculators, estimates, or quote-based pricing.
• Service-specific rules can override the general bandwidth page, especially for storage, CDN, or private connectivity products.
• Dedicated connectivity and CDN paths may have separate pricing models and allowances.

AWS egress pricing: where charges start and where costs spike

AWS is a useful example because the pricing structure is explicit and tiered. The supplied evidence states that the first 100 GB of data egress from Amazon EC2 instances to the public internet is free. After that, data transfer out is charged in tiers, with rates ranging from $0.0900 to $0.0500 per GB depending on monthly volume.

• Outbound data to the public internet is the classic egress scenario.
• Traffic from AWS services to your own data center can also incur data transfer out charges.
• Cross-region and availability zone movements should be checked separately, because not all transfers are priced the same way.
• For teams with multiple services, the surprise often comes from many small transfers adding up across a large footprint.

Azure bandwidth pricing: why the calculator matters

Azure frames bandwidth as data moving in and out of Azure data centers, as well as between Azure data centers. The pricing page explicitly says data transferred out of Azure data centers is covered under bandwidth pricing, but it also notes that estimates are not intended to be final quotes. Pricing can vary by region, offer, date of purchase, and currency exchange rate.

• Use the Azure pricing calculator for the current region and offer.
• Do not assume one region’s outbound rate matches another.
• Be aware that calculated estimates may differ from actual billed amounts.
• Consider whether traffic is routed through CDN, ExpressRoute, or peering, since those paths may be treated separately.

Oracle egress economics and dedicated connectivity

Oracle describes egress as a per-byte fee for data leaving a region, and it notes that networking infrastructure ownership differs across providers and geographies. That matters because cloud bandwidth costs are not just about storage or compute; they are also shaped by the network path and the market where the traffic travels.

• Oracle’s egress framing is region-based rather than a single universal price.
• Pricing varies globally, so cross-region comparisons should be handled carefully.
• Dedicated connectivity can improve privacy, security, and reliability, but it may add substantial cost.
• FastConnect should be evaluated separately from standard internet egress, not blended into the same assumption.

Backblaze B2 as the low-egress benchmark

Backblaze B2 stands out in comparisons because it offers a very different cost profile for outbound transfer. The supplied evidence says B2 users get free egress up to 3x their average monthly storage, then pay just $0.01/GB beyond that threshold. It also states that B2 offers unlimited free egress through partner CDNs and compute services, and that B2 Overdrive includes unlimited free egress to any destination.

• Free egress up to 3x average monthly storage is a strong baseline for many workloads.
• $0.01/GB beyond the threshold is far below many standard cloud transfer rates.
• Partner CDN and compute routes can materially change the economics of delivery.
• B2 Overdrive is positioned for throughput-intensive workloads such as AI/ML, HPC, analytics, and media processing.

Common bill shock scenarios to watch for

• Media downloads and software distribution that generate steady public internet traffic.
• Backup restore events, especially when large datasets need to be pulled back quickly.
• Cross-cloud migrations or cloud-to-on-prem transfers that were not modeled in advance.
• High-volume application delivery without a CDN in front of origin servers.
• Analytics, AI/ML, and HPC pipelines that repeatedly move large datasets between systems.

How to estimate your egress exposure before migration or launch

• Measure monthly outbound GB by workload instead of using one blended estimate.
• Separate internet-bound, cross-region, and on-prem destinations.
• Apply any free-tier allowance before calculating per-GB charges.
• Model best-case, expected, and peak usage to capture real-world spikes.
• Check whether the transfer path is billed as standard egress, CDN delivery, peering, or dedicated connectivity.

Practical ways to reduce cloud transfer costs

• Cache and deliver static assets through a CDN whenever traffic patterns justify it.
• Keep compute and storage in the same region when possible.
• Reduce unnecessary service-to-service chatter, especially in chatty microservice environments.
• Use partner CDNs or bundled transfer paths where they are available and relevant.
• Consider dedicated connectivity only when traffic volume and reliability requirements justify the extra expense.

A durable way to compare providers

When you compare cloud egress fees, the question is not just “Who is cheapest?” It is “Which provider’s transfer model matches my workload?” A platform with a generous free tier may be ideal for bursty delivery, while a quote-based enterprise network path may be better for regulated traffic or predictable large-scale transfers. The most reliable comparison uses actual traffic patterns, destination types, and regional assumptions rather than generic monthly averages.

For teams that are also thinking about architecture choices, bandwidth economics often connect to broader infrastructure trade-offs. That is why transfer pricing should be reviewed alongside performance, reliability, and operational complexity—not after the bill arrives.

Related reading: Optimizing Mobile Backends for Next-Gen Devices: Bandwidth, Sync, and Feature Flags

As providers update free tiers, regional rates, and connectivity options, this guide should be revisited. The biggest savings usually come from catching the expensive transfer path before it becomes a recurring habit.