Cloud Egress Cost Reduction:
Complete Strategy Guide 2026

Why Egress Costs Matter More Than You Think

Cloud egress costs are a persistent source of surprise in enterprise cloud budgets. Unlike compute and storage costs — which are relatively predictable and well-tracked in FinOps tooling — egress costs often accumulate silently across dozens of applications and data flows before appearing as a significant line item. The problem is structural: hyperscalers charge for data leaving their networks because outbound data transfer is one of the most effective forms of vendor lock-in. High egress costs make multi-cloud strategies more expensive and cloud-to-cloud or cloud-to-on-premises migration more costly. This is intentional, and it's negotiable.

This guide is part of our Cloud FinOps negotiation guide. The strategies here cover both the technical dimension (architectural changes that reduce egress volume) and the commercial dimension (contract provisions and negotiation tactics that reduce egress pricing). The most effective egress reduction programmes combine both approaches — architectural changes reduce the volume of data subject to egress charges, while commercial negotiations reduce the per-GB rate for unavoidable egress.

How Egress Pricing Works: AWS, Azure, GCP

Each hyperscaler prices data transfer differently, with distinct structures for different transfer types. Understanding which transfers incur charges — and at what rates — is the foundation of any egress reduction strategy. The following represents standard published rates as of early 2026; negotiated enterprise rates can be significantly lower, particularly for high-volume committed customers.

Transfer Type	AWS	Azure	GCP
Inbound (Internet to Cloud)	Free	Free	Free
Outbound to Internet (first 10TB/mo)	$0.09/GB	$0.087/GB	$0.08/GB
Between regions (same provider)	$0.02–0.09/GB	$0.02/GB	$0.01–0.08/GB
Within same region (cross-AZ)	$0.01/GB	Free	$0.01/GB
Via Direct Connect / ExpressRoute	$0.02/GB	$0.025/GB	$0.02/GB (Cloud Interconnect)
To CDN (CloudFront / CDN / Cloud CDN)	Reduced	Reduced	Reduced

Several important nuances: AWS charges for cross-AZ data transfer even within the same region, which can be a significant hidden cost for microservices architectures with high inter-service communication. Azure generally does not charge for intra-region transfers, giving it a structural cost advantage for certain workload architectures. GCP's inter-region pricing is complex, with rates varying by source and destination region pair — with premium network egress rates applying in some configurations.

Architectural Strategies for Egress Reduction

Architectural changes typically produce the largest and most permanent egress cost reductions. Unlike commercial negotiations — which reduce the per-GB rate for unavoidable transfers — architectural optimisation can eliminate entire categories of egress charges. The following strategies, applied systematically, can reduce egress costs by 40–90% for most enterprise workloads.

1. CDN Offloading for Static and Dynamic Content

Content Delivery Networks (CDNs) — CloudFront for AWS, Azure CDN, or GCP Cloud CDN — serve as a cost-effective buffer between cloud origins and end users. CDN egress rates are substantially lower than direct internet egress rates, and each CDN cache hit eliminates an origin egress charge entirely. For web applications, media distribution, and any workload serving content to geographically distributed users, CDN adoption is typically the single highest-ROI egress reduction action. Enterprises that have not fully migrated their internet-facing content delivery to CDN are routinely overpaying for egress by 50–80%.

2. Cross-AZ Traffic Optimisation (AWS)

AWS's $0.01/GB charge for cross-AZ data transfer — applied in both directions — creates a significant hidden cost in microservices architectures where services in different availability zones communicate heavily. Strategies to reduce cross-AZ traffic include: co-locating tightly coupled services in the same AZ (accepting the redundancy trade-off), implementing regional data caches to reduce cross-AZ read traffic, and restructuring data access patterns to prefer same-AZ replicas. For large microservices deployments, cross-AZ optimisation can reduce annual egress costs by $100K–$1M+ depending on inter-service communication volume.

3. Private Connectivity for On-Premises Integration

Enterprises with significant data exchange between cloud and on-premises systems should evaluate AWS Direct Connect, Azure ExpressRoute, or GCP Cloud Interconnect. These private connectivity services offer data transfer rates of ~$0.02/GB versus $0.09/GB for internet egress — a 78% per-GB reduction. At scale, the savings from private connectivity consistently offset the connection costs for enterprises with more than ~5TB/month of cloud-to-on-premises transfer. Private connectivity also reduces internet egress charges on the origin network, producing a compounding cost reduction across cloud and networking budgets.

4. Data Compression and Format Optimisation

Egress charges are volume-based — reducing the volume of data transferred reduces the charge proportionally. Implementing compression for large data transfers, adopting columnar storage formats (Parquet, ORC) for analytics workloads, and moving from row-based to event-driven data patterns can reduce data transfer volumes by 30–70% for the right workloads. These architectural changes often have dual benefits: they reduce egress charges and improve application performance by reducing the latency of data transfers.

5. Workload Co-location and Regional Optimisation

Enterprises that have evolved their cloud architectures organically often have workloads running in different regions that communicate heavily — incurring cross-region data transfer charges at $0.02–0.09/GB. A systematic review of inter-region data flows, followed by workload co-location where architecturally feasible, can identify significant egress cost reduction opportunities. This review is particularly valuable before long-term commitment negotiations — reducing egress baseline before committing produces lower committed spend levels and therefore lower total commitment costs.

Commercial Levers: Negotiating Egress Rates

Beyond architectural changes, commercial negotiations can directly reduce the per-GB rate for unavoidable egress. This is an area where many enterprises accept published pricing as fixed — when in fact egress rates are negotiable for large-volume customers, particularly in the context of broader hyperscaler commitment negotiations.

Egress Waivers in EDP and MACC Negotiations

Both AWS and Microsoft offer egress fee waivers or credits as part of enterprise commitment negotiations. For AWS, egress waivers are most commonly structured as credits applied to monthly bills for qualifying data transfer. For Microsoft Azure, egress fee caps and waivers can be incorporated into MACC terms. Google Cloud has historically been the most willing to negotiate data transfer rates in custom pricing agreements. The key lever: presenting your total egress spend alongside your commitment negotiation, and requesting specific egress pricing treatment as a component of the overall deal. Enterprises that treat egress as a separate conversation from their commitment negotiation consistently achieve worse outcomes than those who bundle it into the broader commercial discussion.

High-Volume Internet Egress Discounts

All three major hyperscalers publish tiered internet egress rates — lower per-GB rates for higher monthly volumes. AWS, for example, applies lower rates for monthly egress above 10TB and 150TB. However, published tier pricing is not the ceiling of available discounts for enterprise customers. Enterprises with more than $100K/month in egress spend regularly negotiate custom egress pricing below the published tiers through direct account team negotiations — particularly when the negotiation is framed around total relationship value and commitment.

Cloud Egress as Contract Leverage

High egress costs represent implicit switching costs — they make migration to competing clouds or on-premises more expensive. Ironically, this same switching-cost dynamic can be used as negotiation leverage: demonstrating to a hyperscaler that you have implemented architectures that could reduce your egress exposure (through multi-cloud routing, edge computing, or CDN strategies) changes the commercial dynamic. A buyer who is genuinely capable of reducing egress dependency is a more credible negotiating counterpart than one who accepts egress costs as unavoidable.

The Regulatory and Policy Context

Cloud egress costs have attracted significant regulatory attention in Europe and are increasingly on the agenda of procurement regulators globally. The EU's Data Act (effective 2025) includes provisions aimed at reducing switching costs — including egress fees — for cloud customers. Several EU member states have incorporated cloud portability and egress cost reduction requirements into public sector procurement rules. For enterprises subject to EU data governance requirements, these regulatory developments create additional leverage in hyperscaler negotiations: citing regulatory compliance requirements around data portability as a reason for requesting reduced or waived egress fees is increasingly effective, particularly with European offices of AWS, Azure, and GCP.

The UK Cloud Market Study by the Competition and Markets Authority (CMA) has also highlighted egress costs as a competition concern, and has recommended pricing reforms. While regulatory outcomes are uncertain, the trajectory of regulatory pressure is clearly toward reduced egress fees for enterprise buyers — and forward-looking hyperscalers have been increasingly willing to negotiate egress terms proactively to stay ahead of regulatory mandates.

FinOps Tooling for Egress Visibility

Effective egress management requires visibility into egress patterns at a granular level. The native cost explorer tools provided by AWS, Azure, and GCP provide egress data at a service and region level, but typically lack the granularity to identify specific application-level data flows driving high egress. Third-party FinOps platforms — CAST AI, Apptio Cloudability, CloudHealth, and others — provide more detailed egress analysis and can identify the specific service-to-service and service-to-internet data flows that account for the majority of egress spend.

Our recommended approach for a comprehensive egress review as part of a broader cloud cost optimisation engagement: analyse 90 days of detailed billing data at the resource level; map high-egress resources to specific application teams; assess whether architectural changes are feasible for the top 5–10 egress contributors; then use the resulting analysis as input to hyperscaler negotiations. This approach typically identifies 40–60% egress reduction opportunities within a 90-day engagement.