[The short URL for this blog post is cdnmarket.com] Since my last CDN market sizing post in 2023, a lot has changed. Edgio went bankrupt, StackPath closed down, and Lumen exited the CDN business, with Akamai acquiring select contracts (not assets) from all three companies. While the pool of CDN vendors shrank from 2023 to 2026, the CDN industry’s total revenue growth for delivery services was flat or down by low single digits (-2%). Traffic shifted among vendors due to consolidation, but no market catalyst accelerated bit-delivery growth over the past few years.
What we initially saw post-COVID with streaming services optimizing their encoding and bitrate ladders accelerated, with many delivering the same or better video quality with fewer bits. 4K streaming consumption didn’t grow, with the industry defaulting to 1080p HDR for most live events. Some of the largest OTT platforms that had 4-6 vendors in their multi-CDN strategy pulled back to 2-3 vendors and openly discussed not seeing benefits from using more than that. In the past two years, many customers, especially at the mid-tier level, have been unwilling to make year-long, high-volume commitments, with traffic allocations within a multi-CDN stack becoming significantly more dynamic, occurring at the country or even single-ISP level.
Some large-scale live events, while not major revenue drivers, moved from streaming services that use third-party CDNs to platforms such as Netflix and YouTube, which don’t use third-party vendors for video delivery. Amazon, one of the largest customers of third-party CDNs for delivering Prime Video content, brought more of its delivery traffic in-house to CloudFront over the three-year period. TikTok, a major CDN customer across multiple vendors, also shifted some traffic distribution and relied more on its own DIY CDN, lowering pricing in its traffic reallocations with CDN vendors.
Lower overall bitrates, combined with OTT services already at scale in the U.S., resulted in fewer new subscribers and slower bit growth for CDNs. The hype around 4K, AR, VR and other formats that were supposed to drive CDN service usage at scale was never real and didn’t materialize. Wall Street’s thinking that the agentic use of the internet and web applications would “drive CDN traffic higher” demonstrated a lack of understanding of delivery technology, the scale of traffic requests, and how content is delivered. Agentic traffic refers to traffic generated by autonomous AI agents. AI agents are not watching videos from streaming services or downloading large software files, which account for the largest majority of the bit volume for CDN vendors.
As in previous years, a major challenge for the CDN industry is that a small number of customers account for a disproportionately large share of total revenue. When any large customer changes their content strategy, merges with other services, or shifts their DIY strategy, the impact is felt by CDN vendors. By my estimate, fewer than 50 CDN customers account for 75% of the revenue generated by third-party CDNs. Over the years, we’ve had multiple data points from Akamai, Fastly and Limelight Networks to support this.
In Q1 2023, fewer than 20 customers out of nearly 900 accounted for approximately 75% of Limelight’s total revenue. Notably, two customers, Amazon and Sony, accounted for approximately 42% of their total revenue, and seven customers combined accounted for just over $181 million. Akamai has not broken out its customer split the same way, but disclosed for many years that, when it broke out revenue under the “media” bucket, six customers accounted for 18% of its media revenue at the high point. In Q1 of 2026, Fastly reported that its top 10 customers accounted for 34% of the company’s total revenue, not just revenue generated from delivery services.
Based on my regular off-the-record conversations with many customers and CDN vendors, as well as public data points on traffic volumes and filings, here is an incomplete list of the largest CDN customers: Amazon (Prime Video), Disney, NBCU, Paramount Skydance, Warner Bros. Discovery, TikTok, Microsoft, Apple, Sony Interactive Entertainment, Nintendo, Activision Blizzard, Electronic Arts, Riot Games, Valve, Take-Two Interactive, Roblox, Ubisoft, Roku, NFL, ESL FACEIT Group, Spotify, Reliance Industries, JioHotstar, Snap, The Times Group, X, and Reddit, amongst others.
CDN Pricing Trends and Impacts
I’ll do a separate post soon with an update on CDN pricing, but for those who think pricing automatically falls by X% each year, those days are over. Full stop. There is no “race to the bottom” as many suggest. For the largest customers, the rate of price declines over the two years was stable, but even a 5% price reduction still affects many vendors with low margins in delivery services and high capex. The rising costs of hardware, power and other infrastructure-related services, which increased significantly in the second half of 2025, forced Akamai to add a 3% monthly surcharge for all clients, starting in Q2 of this year. Do not expect CDN pricing to decline in the coming years; it will slightly tick up for most.
With CDN vendors facing rising infrastructure-related costs, lower overall bitrate delivery, and OTT services already at scale in the U.S., CDNs are not making capacity investments without guaranteed bandwidth or revenue commitments. Gone are the days of “build it, and the traffic will come.” Pricing in South America averages 2.8x per GB compared to North America, and capacity in Korea is almost nonexistent, at 10x the North American price. CDNs are being squeezed by ISPs, and the cost of chips, storage and power has all gone up. Some CDNs are no longer absorbing egress charges or issuing egress credits, and many are not building out additional capacity in regions where the market opportunity is not large enough to guarantee a return on the capex. Case in point: much of the traffic flowing into LATAM still comes from Miami.
I’ve also seen deals where customers allow CDNs to deprioritize their traffic in exchange for a lower rate and a lower SLA. Data sovereignty laws in Europe are also causing issues for some CDNs, with customers demanding that CDNs serve content from a cache in the country, forcing some CDNs to pass on select RFPs. While CDN pricing can be slightly discounted for deals involving compute and services like WAF, don’t listen to anyone who continues to post on LinkedIn that some CDN vendors are “giving away” delivery to get higher-margin cloud security services. That is 100% inaccurate.
Discounts for bundling do occur, but I see multiple CDNs passing on deals that are too low and will not price below their cost. As a recent example of pricing for a contract in the tens of millions annually, Akamai, Fastly, CloudFront, Google Media CDN, and CDN77 all quoted pricing in an RFP within $0.0001 of each other per GB for North American delivery. No CDN vendor is undercutting competitors by a large margin. It’s not happening. Those who post on LinkedIn suggesting delivery is a “race to the bottom” and suggest pricing “falls by 20% each year” have no idea what they are talking about. They don’t see contracts, per-region pricing, commitment levels, SLA guarantees, etc., because if they did, they would know it’s all about margins. Too many people with no real insight into the CDN market continue to share theories and opinions on LinkedIn disguised as facts, but facts they are not.
CDN Market Sizing Definitions
When discussing market sizing numbers for any product or service, only three things matter: the methodology, definitions and facts. Many incorrectly use public revenue figures from vendors, changing the terms the vendors use, and then quote the company incorrectly. Others make up numbers based on rumors and then state them as facts, and all CDN reports I see don’t clearly define which CDN services are covered and which are not included. Do not buy any CDN report. They are all garbage, every single one of them. I will answer your questions, free of charge, and if I don’t have the answer, I’ll track it down for you from those who do. (917-523-4562, dan@danrayburn.com)
While my post mentions video as one CDN use case, my CDN market sizing numbers include delivery of large and small objects, as well as streaming. My figures do not include revenue from services tied to cloud security (WAF, DDoS, etc.), compute, bare-metal and media workflow platforms like Uplynk, Brightcove, AWS Media Services, MediaKind and the like. I estimate the media platforms market, excluding delivery services and hardware, generated approximately $2.5 billion in revenue in 2025.
No vendor breaks out revenue from CDN services solely for video delivery. Only once in the past six years has Akamai disclosed what percentage of total bits delivered on its network came from video. Bits, mind you, not revenue. At their 2021 investor summit, the company said 57% of all bits delivered under their “Edge Delivery Traffic” bucket were OTT/video in fiscal year 2020. While many CDN vendors have provided me with estimates of how much revenue they believe comes from video delivery, I am not releasing any vendor-specific numbers. All I will disclose is that if you add up all of the revenue CDN vendors get for bit delivery, less than half comes from video content specifically.
The terminology vendors use to report CDN or delivery revenue numbers to Wall Street is very specific for legal reasons and can also change over time. Many continue to get this wrong, year after year. For the last few years, Akamai has broken out company revenue into three buckets: “Delivery,” “Security,” and “Compute.” Not CDN and not video. Recently, Akamai changed how it reports revenue across buckets, grouping them into “Delivery and other cloud applications,” “Cloud Infrastructure Services,” and “Security.” Fastly uses the terms “Network Services” (solutions designed to improve the performance of websites, apps, APIs, and digital media), “Security” (products designed to protect websites, apps, APIs, and users) and “other” (emerging product offerings which include compute and observability products).
It’s essential to understand these distinctions to accurately interpret the revenue figures and to understand that they are not always an apples-to-apples comparison. For example, Akamai includes revenue from API traffic under its security bucket, while Fastly includes that traffic under its Network Services bucket, which includes delivery as one component of the grouping.
Carrier and ISP Platform Vendors
While not the same as third-party CDNs that rent their infrastructure to content owners, vendors selling their CDN platforms to telcos and ISPs to deploy capacity within their private networks saw overall revenue decline over the past two years, with Qwilt in particular nearly going out of business. Vendors who have been pitching a revenue-share model to ISPs have finally realized it doesn’t work and should have moved away from it years ago. Telcos and ISPs deploy caches from vendors or content owners, or build their own, to do one thing: reduce transit costs. The majority are not trying to monetize the traffic on their network.
We knew many years ago that trying to convince telcos and ISPs to enter the delivery business and sell CDN services to content owners didn’t work, as evidenced by the failed attempts by Comcast, Verivue, Ericsson, and others. Approximately $150 million was generated in 2025 from vendors offering CDN platforms to operators and carriers, including ATEME, Broadpeak, Gcore, Jet-Stream, Netskrt, Synamedia, Qwilt, Varnish Software, Vecima, Velocix and others. My $150 million figure discounts these vendors’ total revenue by the percentage I believe is allocated to their CDN product, which, for many, is their lowest-contributing product offering.
CDN Vendor Revenue Breakdown
- Akamai reported $1.25 billion in delivery revenue for 2025, down 5% from the prior year. That number included revenue from contracts acquired from Edgio, StackPath and Lumen, which, in the absence of those deals, would have caused Akamai’s delivery revenue to fall close to 9% in 2025. Those posting online, suggesting that the recently announced $1.8 billion committed revenue contract is for its Cloud Infrastructure Services, rumored to be from Anthropic, will help “save a legacy CDN vendor,” are idiotic. Akamai’s security segment became the company’s largest revenue generator in Q1 2023, and in 2025, 70% of Akamai’s revenue came from services outside of its delivery business. Akamai did a great job of diversifying its revenue stream away from delivery and media many years ago, a feat no other pure-CDN vendor has matched to date.
- Amazon doesn’t disclose CloudFront revenue, but I estimate it grew from $1.1 billion in 2023 to $1.5 billion in 2025. Anyone who tracks traffic share and delivery contracts has seen how aggressively Amazon has pursued a larger share of delivery traffic over the past few years. I am not disclosing the source of my estimate, but note that at one time Amazon published the number of paying customers it had on CloudFront and its growth rate. It is important to note that a large percentage of CloudFront’s revenue, I estimate close to 50%, would not be comparable to Akamai’s or Fastly’s delivery business, which doesn’t take on customers paying $100 a month. Amazon has many customers too small for larger CDNs based on their ARR, and I hear that CloudFront now has well over 100,000 paying customers.
- CDN77 disclosed that its 2025 revenue was $218 million, up from $145 million in 2023, and that it was cash-flow-positive. While the company’s revenue is lower than that of some larger CDN vendors, its CDN business grew at a CAGR of over 40% between 2020 and 2025. To date, they are only the second CDN vendor in history (see cdnlist.com) to reach $200+ million in delivery revenue and be cash-flow-positive. Read that again. An aspect that makes CDN77 unique in terms of revenue is that, unlike other CDN vendors, a large majority of its revenue comes from storing, processing and delivering video. While video is CDN77’s focus, I expect the company to expand its CDN business by offering large-object delivery and other non-CDN services, such as compute, later this year.
- Cloudflare reported $2.16 billion in revenue for 2025, of which I am counting $500 million towards my market sizing numbers. However, it is important to call out that Cloudflare is the wildcard in sizing the market, since the majority of what they deliver is neither video nor large objects. The market sizing numbers I’ve calculated could fluctuate by a few hundred million, depending on how others define the CDN or delivery business. Cloudflare offers many cloud security and DNS services that do not fall under my definition of delivery, and the company does very little in streaming. Cloudflare has no interest in competing with other CDNs in live streaming, software downloads or large-object delivery, so while they are a CDN, I don’t count them as a primary CDN in the market for the services talked about in this post. You should discount my numbers as you see fit, based on how you want to segment out Cloudflare’s involvement in the delivery market.
- Fastly reported $624 million in total revenue for 2025, of which $477.8 million was allocated under its “Network Services” bucket, making up 76% of its total revenue. I am discounting that number slightly to account for revenue from non-delivery services, and I am using $400 million for my market sizing estimate.
- Google‘s CDNs, Google Cloud and Google Media CDN, I estimate, had $175 million in revenue in 2025. In April 2025, Google allowed me to disclose in my blog post that Google Media CDN had grown to over 100 Tbps of egress capacity and that customers, including MLB and Warner Bros. Discovery, were using Google as part of their multi-CDN approach. Earlier this year, Google shared additional data points about its business, noting that it supported live streaming events, including the Super Bowl, the FIFA World Cup, and the IPL. Google’s initial market entry with its Media CDN offering was focused on its network foundation. The changes over the last year, however, indicate a shift toward addressing more specific, operational, and architectural challenges faced by large-scale broadcasters.
Outside of the top six prominent CDN vendors, no other vendor has reached $100 million in annual run-rate revenue. While Comcast, MainStreaming, Medianova, CacheFly, Qwilt, Wowza, Bunny.net and others don’t publicly disclose revenue numbers, most have provided me with numbers and guidance, off the record or under an NDA. Netskrt allowed me to publish their 2025 revenue numbers, and Gcore’s numbers are available in public documents. Based on that data, grouping them together would result in an additional $400 million in delivery revenue for the CDN market in 2025.
Dozens of smaller regional CDNs target SMB customers and low-traffic customers, with an estimated $500 million in combined annual revenue. It is worth noting that although all these vendors fall under the delivery revenue category, they don’t all compete. Not all delivery revenue is comparable across vendors, and many of these smaller CDNs are also offering services tied to hosting, storage, etc.
2025 Market Size
For 2023, I estimated the total revenue of all vendors offering delivery services, excluding vendors in China, was $4.95 billion. That number could fluctuate by a few hundred million, depending on how much of a company’s total income is allocated to “delivery.” The number includes vendors listed on cdnlist.com, smaller regional CDNs, and revenue from vendors that offer CDN platforms to telcos and ISPs. From 2026 to 2028, I expect the overall CDN market revenue to grow between 2-3%, noting that if prices go up, that doesn’t mean the overall market demand grows, only the size of the total addressable market.
If you have any comments or questions about this post or the topic, you can leave them in the comments section here on LinkedIn or contact me directly at dan@danrayburn.com or 917-523-4562. I am aiming to publish my post on CDN pricing trends soon, including data from an updated Q2 survey.
Note: I have never bought, sold, or traded shares in any public CDN, and even in my managed portfolios, Akamai, Fastly and Cloudflare are excluded. Limelight Networks and Edgio were also excluded when they were public companies.
