Learn
Best Video Streaming CDN in 2026? 7 Providers Compared With Real Performance Data
Best CDN for Video Streaming in 2026: 7 Providers Compared A single rebuffer event at the two-second mark costs you 8% ...
A mid-market SaaS platform running 50 TB/month through Fastly can see its effective per-GB rate swing from $0.01 to north of $0.04 depending on traffic mix, region split, and a handful of line items that never appear on the pricing page headline. If you are evaluating Fastly CDN pricing in 2026, the sticker rate is the least interesting number. The interesting numbers are the seven multipliers hiding beneath it. This article gives you the exact cost components, a region-by-region bandwidth table current as of Q2 2026, and a workload-profile decision matrix you will not find in Fastly's docs or any competing top-10 result.

Fastly bills on two primary axes: bandwidth delivered from the edge and the number of HTTP/HTTPS requests served. Both dimensions are tiered by geography, which means a single global traffic profile produces a blended rate that is almost never the number you saw on the pricing calculator. As of Q2 2026, Fastly's published on-demand bandwidth starts at approximately $0.08/GB for North America and Europe, rises to roughly $0.12/GB across most of Asia-Pacific, and can reach $0.19/GB or higher in South America and parts of Africa. Request pricing adds $0.0075 per 10,000 HTTPS requests in NA/EU and scales up in other regions.
These base rates only tell half the story. Below are the seven cost multipliers that consistently surprise teams after the first full billing cycle.
Fastly segments the globe into billing regions, and the spread is not trivial. Serving 10 TB from Sydney costs roughly 50% more than serving the same volume from Ashburn. Teams that model costs using a single global average get burned the moment APAC or LATAM traffic ticks up. If your user base is shifting toward Southeast Asia — a trend accelerating through 2026 — your Fastly bandwidth pricing will drift upward even if total volume stays flat.
A modern SPA or media-heavy page can fire 80–200 sub-requests per page load. At Fastly's request pricing tiers, a site doing 100M page views/month with an average of 120 requests per page is looking at roughly 12 billion requests. Even at the lowest published tier that is a five-figure monthly line item for requests alone, before bandwidth. Fastly request pricing is often under-modeled because teams focus on GB delivered.
Origin shielding reduces origin load, but every shield-to-edge fetch still counts as internal bandwidth and, depending on contract terms, can be billed. In high-cardinality catalogs (video libraries, large e-commerce SKU sets) shield hit rates below 70% mean a meaningful share of traffic is double-billed: once from origin to shield, once from shield to edge. Monitor your shield efficiency ratio monthly.
Fastly includes shared TLS on its default domain, but dedicated TLS with custom certificates — a non-negotiable for any brand-conscious production deployment — adds per-domain charges. As of 2026, dedicated TLS runs in the range of $100/month per SAN slot, and wildcard certificates cost more. Organizations operating multi-brand or white-label platforms with dozens of custom domains can accumulate four-figure monthly TLS costs.
Fastly's real-time log streaming is a genuine differentiator, but the data has to go somewhere. Streaming to an external endpoint (Datadog, Splunk, S3) generates egress volume that can be substantial for high-request-rate services. This egress is often metered separately and frequently overlooked during capacity planning.
Instant purge is one of Fastly's strongest technical features. But teams running high-frequency programmatic purges — common in news, financial data, and inventory systems — may hit rate limits that require upgraded API tiers or custom contract terms. The operational cost is not always dollars; it can be engineering time spent building surrogate-key strategies to stay within purge budgets.
Compute@Edge workloads are billed on invocations, duration, and memory. A seemingly lightweight edge function executing on every request at 500M requests/month can generate a bill that dwarfs the underlying bandwidth cost. In 2026, Fastly's compute billing granularity is per-50ms execution window with a memory multiplier. Profile your edge functions carefully; cold-start variance alone can shift monthly compute spend by 20%.
Fastly offers on-demand, committed-use, and enterprise contract pricing. The on-demand rates listed above are the ceiling. Committed contracts (typically 12 months, minimum ~$2,500/month) unlock lower per-GB and per-request rates, but they lock you into a minimum spend regardless of actual usage. Enterprise agreements are opaque and negotiable. The decision depends on traffic predictability:
| Workload Profile | Recommended Tier | Why |
|---|---|---|
| Steady-state, NA/EU-heavy, >80% cache hit ratio | Committed contract | Predictable volume makes committed minimums safe; per-GB discount offsets lock-in risk. |
| Bursty (events, launches), global, high request fanout | On-demand or multi-CDN | Burst headroom without overpaying during troughs; combine with a cost-efficient secondary CDN. |
| High-volume media/video, >100 TB/month | Enterprise negotiated or alternative provider | At this scale, per-GB rate dominates total cost; every $0.005 delta is $500+/month per 100 TB. |
This matrix is the section missing from every existing Fastly pricing analysis. Instead of asking "how much does Fastly cost per GB," ask "given my traffic shape, which cost levers dominate and which provider minimizes them?"
| Criterion | Fastly Strong Fit | Fastly Weak Fit |
|---|---|---|
| Purge latency requirement | <150 ms global purge needed | Purge frequency low; latency tolerance >5 s |
| Edge compute complexity | Wasm-based logic at the edge is core to architecture | Static asset delivery with minimal edge logic |
| Monthly volume | <50 TB, NA/EU concentrated | >100 TB, globally distributed |
| Budget sensitivity | Performance justifies premium | Per-GB cost is the primary constraint |
For workloads that land in the "weak fit" column — high-volume, globally distributed delivery where per-GB economics dominate — the math often points toward providers built for cost efficiency at scale. BlazingCDN is worth benchmarking here: pricing starts at $0.004/GB ($4/TB) for volumes up to 25 TB and drops to $0.002/GB at the 2 PB tier. That is a 2–4x cost reduction compared to Fastly's on-demand rates for equivalent geography, with uptime and fault tolerance on par with CloudFront. Enterprises including Sony use BlazingCDN for high-volume delivery where predictable, flat-rate billing eliminates the hidden-cost problem entirely.
Run this audit before your next contract renewal:
On-demand, Fastly charges approximately $0.08/GB in North America and Europe, scaling to $0.12–$0.19/GB in Asia-Pacific, South America, and Africa as of Q2 2026. Committed contracts lower these rates by 15–30% depending on volume and term length.
Fastly bills per 10,000 requests, with HTTPS requests priced at roughly $0.0075/10k in NA/EU. The rate increases in other billing regions. High-fanout SPAs and API-heavy applications accumulate request charges quickly — often faster than bandwidth charges.
It can. While shielding reduces origin load, every shield-to-edge transfer is billed as delivered bandwidth. If your shield hit ratio is below 70–75%, the incremental bandwidth cost may exceed the origin savings. Audit shield efficiency before assuming net benefit.
Shared TLS on Fastly's default domain is included. Dedicated TLS with custom certificates costs approximately $100/month per SAN slot as of 2026. Wildcard certificates carry a higher rate. Multi-brand deployments should model this cost explicitly.
Fastly divides delivery into billing regions — NA, EU, APAC, South America, Africa, and others — each with its own per-GB rate. Your invoice reflects the blended cost across all regions, weighted by actual bytes delivered in each. Shifting user demographics directly shift your effective rate.
Yes. Increase cache hit ratios through better surrogate-key design and longer TTLs. Consolidate TLS domains. Reduce request fanout at the application layer. Move analytics log streaming to sampled mode if full fidelity is not required. These optimizations can cut 15–25% from a typical bill.
Pull your last three Fastly invoices and break out the per-region, per-request, and per-compute line items into a spreadsheet. Calculate your blended effective rate and compare it against the on-demand published rate for your dominant billing region. If the delta exceeds 20%, you have hidden costs worth chasing. Then run the same workload estimate through two alternative providers and compare. The number that matters is not the headline per-GB rate — it is the fully loaded cost at your actual traffic shape. Start there.
Learn
Best CDN for Video Streaming in 2026: 7 Providers Compared A single rebuffer event at the two-second mark costs you 8% ...
Learn
Video CDN Providers Compared: BlazingCDN vs Cloudflare vs Akamai for OTT If you are choosing a video CDN for an OTT ...
Learn
Video CDN Pricing Explained: How to Stop Overpaying for Streaming Bandwidth Video already accounts for 38% of total ...