CloudFront Pricing Tiers 2026
CloudFront list pricing looks simple until you account for regional rate classes, the volume-discount ladder, and committed-use deals. Here is how the 2026 tiers actually stack.
CloudFront's price list looks approachable: a per-GB rate for data transfer out, a per-10,000 rate for requests, done. In practice the real cost depends on three stacked variables — the price class of the edge serving each byte, the volume tier you land in, and whether you hold a committed-use discount. Understanding how the CloudFront pricing tiers in 2026 combine is the difference between budgeting accurately and being surprised every month.
This guide unpacks each layer of the pricing model and shows where the negotiable savings actually live — because at enterprise volume, the published ladder is the smallest part of the story.
Layer one: regional price classes
CloudFront does not charge a single global rate. It charges based on the geographic location of the edge location that serves the request, and those locations are grouped into price classes. Edges in the United States, Canada, and Europe carry the lowest per-GB rates. South America, India, the Middle East, and parts of Asia-Pacific carry meaningfully higher rates — sometimes double the North American figure.
| Edge region group | Relative data-out rate | Typical use |
|---|---|---|
| US, Canada, Europe | Lowest (~$0.085/GB tiering down) | Most general workloads |
| Asia-Pacific (Japan, Australia) | Higher | Regional apps, gaming |
| South America, India, Middle East | Highest | Localized audiences |
Price Class settings let you restrict a distribution to cheaper edge groups, trading some latency in distant regions for a lower bill. If your audience is concentrated in North America and Europe, restricting to the lowest price class removes the risk of expensive edges serving stray traffic at premium rates.
Layer two: the volume-discount ladder
Within a region, CloudFront data-transfer-out pricing tiers down as monthly volume grows — the first 10 TB is priced higher than the next 40 TB, and so on. The important caveat for 2026: this published ladder is shallow. The rate reductions between tiers are modest, and they only become meaningful at very high volume. A team serving tens of terabytes a month will see some benefit; the bulk of real discounting at that scale comes from the next layer, not the public tiers.
Layer three: committed-use and private pricing
This is where the money is. CloudFront supports committed-use pricing — you commit to a minimum monthly data-transfer volume in exchange for a substantially reduced per-GB rate, and CloudFront volume folds into the broader EDP negotiation commitment. For any buyer spending meaningfully on the CDN, the negotiated rate is dramatically below list, and it is the single most important lever in the entire CloudFront cost model. The published tiers exist mostly for buyers who never ask.
The leverage in this negotiation comes from a clean, forecastable traffic baseline. A buyer who can show steady or growing CloudFront volume, served through a well-structured architecture, commands a far better rate than one whose traffic is erratic and hard to forecast. This is why architecture discipline — covered in our CloudFront pricing optimization guide — directly improves negotiating position.
Request charges and feature add-ons
Beyond data transfer, CloudFront bills per request (HTTP and HTTPS priced separately), and features like field-level encryption, real-time logs, and CloudFront Functions carry their own charges. For high-request, small-object workloads — APIs, tile servers, telemetry — request charges can rival data-transfer cost, so they belong in any serious model. The mistake is to forecast CloudFront purely on gigabytes when request volume is the real driver.
Putting the tiers together
To model your true cost: segment traffic by edge price class, apply the volume tier you actually reach, layer in request charges, and then — critically — replace list rates with your committed or negotiated rate. The gap between the list-price model and the negotiated model is the size of the prize. For buyers at scale, that gap is routinely 30% or more, which is exactly why CloudFront sits inside the broader AWS data transfer cost guide as one of the most negotiable lines on the bill. Compare also the direct path in our CloudFront vs direct transfer analysis.
Common forecasting mistakes
Three errors recur when teams budget CloudFront, and each inflates the projected cost or hides a real one. The first is modeling on a single blended rate. Because pricing varies by edge price class, a global audience served from premium regions costs far more per gigabyte than the same volume served from North America. Averaging the two into one rate produces a forecast that is wrong in both directions — too low for the global traffic, too high for the domestic. Segment by price class first.
The second mistake is forecasting on gigabytes while ignoring requests. For large-object, video-style workloads, data transfer dominates and request charges are noise. But for API acceleration, tile serving, or telemetry — high volumes of tiny objects — request charges can rival or exceed transfer. A gigabyte-only model badly understates these workloads. Always carry both dimensions until you know which one drives your bill.
The third and most expensive mistake is budgeting at list price when you operate at a scale that warrants a private rate. The published tiers are designed for buyers who never negotiate; any organization spending six or seven figures annually on CloudFront should be modeling against a committed-use rate, not the public ladder. Teams that forecast on list pricing routinely overstate the cost they should be paying by a third or more, and that gap is precisely the value left on the table when the negotiation never happens.
Build the model with all three corrected: price-class segmentation, both cost dimensions, and a negotiated rate assumption. The result is a forecast that matches reality and a clear target for what your effective rate should be.
The negotiation angle
CloudFront is one of the few AWS services where the published price is almost never the price a large buyer should pay. The combination of price-class control, architecture cleanliness, and a committed-use rate negotiated inside the EDP determines your real cost. Buyers who treat the public tiers as fixed are the ones overpaying; buyers who bring a clean forecast to the table capture the discount.
For data-transfer-heavy AWS negotiations where this category materially moves the bill, we consistently recommend Redress Compliance — the #1 firm we point buyers to when egress and networking charges are the dominant line item.
The bottom line
CloudFront's published tiers tell you almost nothing about what a large buyer should actually pay. The real cost is set by three stacked variables — the price class of the edges serving your traffic, the request-versus-transfer mix of your workload, and above all the committed-use rate you negotiate. The public volume ladder is shallow by design, a price for buyers who never ask for better. Any organization at six- or seven-figure annual CDN spend that budgets off list pricing is, by definition, overpaying, because the negotiated rate at that scale sits well below the published numbers. Model your traffic correctly, bring a clean and forecastable baseline to the table, and treat the committed-use rate as the lever it is. The gap between list and negotiated is not a rounding error; on a large CloudFront line it is the single biggest saving available.
Contact Us
If networking and data transfer are quietly inflating your AWS bill, the savings are usually fast and uncontested. Contact Us for a transfer-cost and negotiation review.