Morning, good afternoon, and good evening. I'm Tom Barth, head of investor relations, and on behalf of our over 8,000 employees worldwide, I want to welcome you to Akamai's 2021 Investor Summit. We are grateful that you have chosen to spend your valuable time with us. As you can see, we have a full agenda for you today, as we have a lot to talk about. There are a couple of planned short breaks, and after all the presentations, we will conclude today's summit with an extended Q&A session. While we will hold your questions to the Q&A, you can submit your questions at any time throughout the morning by entering it in the questions box, and we will incorporate them into the Q&A session.
We hope you walk away from today with a very clear picture of the market forces and opportunities in front of Akamai and how we intend to capitalize on them, as well as an in-depth look at Akamai's financial profile. Before we get started, I'm sure you will all be grateful to hear that I don't intend to read all of this slide. As you know, it's not a summit without our best friend, Safe Harbor. Please note that today's presentations and comments may include forward-looking statements. These forward-looking statements are subject to risk and uncertainties and involve a number of factors that could cause actual results to differ materially from those expressed or implied by such statements, particularly statements about our target model and future financial performance.
Those factors are listed here, but include uncertainty stemming from economic conditions, the COVID-19 pandemic, any impact from unexpected geopolitical developments, M&A, and other factors. Additional information concerning these factors are contained in Akamai's filings with the SEC, including our own annual report on Form 10-K and quarterly reports on Form 10-Q. The forward-looking statements included in these presentations and comments represent the company's view on February 25th, 2021. Akamai disclaims any obligation to update these statements to further reflect events and circumstances. As a reminder, we'll be referring to some non-GAAP financial metrics during today's presentations. A detailed reconciliation of GAAP and non-GAAP metrics can be found under the financial portion of the investor relations section of akamai.com. Okay, enough of all that. Let's get started.
It started with an idea over 20 years ago to solve the greatest problem of the early internet, the World Wide Wait. That idea became Akamai, and we unlocked the power of the internet to connect clouds, data, and people instantly, almost magically. Akamai proved that the shortest distance between two points is not the straightest line. It's the smartest line. In today's fast, vast, and volatile digital world, the core and the cloud can no longer keep up. Many clouds, new costs, bigger threats. Billions of people, devices, and streams of data are fighting for capacity and speed. Technology's center of gravity is shifting from the cloud to the edge. Akamai is that edge, the world's largest and most powerful edge platform, surrounding and protecting everything from the enterprise to the cloud, keeping experiences closer to users and threats farther away.
We built something no one else has, an edge platform without trade-offs. A place where innovation, security, and technology all work together. An edge that's transforming the way we live, work, and play. Live streaming the highest profile events to the largest global audiences and stopping the most dangerous cyberattacks in history. When COVID-19 forced the world's activities online, our technology and ingenuity were put to the test. The internet became a lifeline. As more people gripped that lifeline tighter and longer, security risks skyrocketed. Akamai was ready for anything. We helped companies worldwide connect and protect remote workforces practically overnight. We leveraged our scale and support for healthcare, education, entertainment, e-commerce, financial services, and social networks. Amid unprecedented demand, we continued making the internet fast, reliable, and secure. Through it all, our customers know that when a new threat appears, they're safe from it.
When a new opportunity arrives, they're ready to capitalize on it. Akamai is the edge without compromise, as agile and secure as this digital world demands, so businesses can adapt and keep moving, whether they're solving this morning's challenges or preparing for a million possible futures.
Hello, and thanks for joining us. We appreciate having this time with you to talk about our business and how we see it evolving over the next several years. As Tom Barth mentioned, you'll be hearing from several of our key executives today. We'll be happy to take your questions following the presentations. As most of you know, Akamai sells services to the world's leading enterprises that enable and protect their business online. Our mission is to make the internet fast, intelligent, and secure. We accomplish this mission using our intelligent edge platform. Our edge platform is the largest and most trusted cloud delivery platform in the world. We operate at a massive scale with over 340,000 servers handling trillions of requests per day and well over 100 Tb per second of concurrent traffic around the clock.
These are enormous numbers, but what's even more important than our scale is where the servers are located. They're at the edge of the internet in over 4,000 different locations, and inside over 1,400 different networks. Not peering with these networks, but inside 1,400 networks, well past congested peering points. We're in over 900 cities and 135 countries, close to where all the end users are on the internet. A lot of companies talk about being at the edge today, but no other company comes anywhere close to this kind of distribution and proximity to end users. That gives Akamai a big advantage over the competition. I think it's also important to understand what all these servers are doing. Of course, they distribute popular content like movies and gaming software.
They also work together as an intelligent fabric to mitigate and minimize network congestion, route around trouble spots on the internet, and to accelerate access to content that is typically not cached locally, like your bank balance or shopping cart. Operators very much like Akamai servers in their network because we can help them avoid overload by diverting traffic away from areas with congestion. We improve the performance for their subscribers for all the content we serve, which is a lot of the content on the internet. We see so much of the internet traffic, our servers also collect a vast amount of data that we can provide to our customers in real time, so they can optimize the business value of their sites and apps.
The same Akamai Edge server that's doing all these things is also providing security to protect websites, apps, enterprise communications, and data. In all, a typical Akamai Edge server is performing dozens of functions on behalf of thousands of customers at any given time. Akamai Edge servers are also programmable by our customers and third parties using our V8 JavaScript engine that's been optimized to respond instantly to requests to run an app. We handled over 300 billion such requests in Q4, and we supported over 300 trillion API requests overall in 2020. Bobby Blumofe is going to talk a lot more about our Edge platform later today, and before I move on, I think it makes sense to say a few words about what the Edge really is. The Edge is the part of the internet that's closest to end users.
This is also often called the last mile, and it sits between the cloud data centers in the core of the internet and the billions of devices that are in homes and offices or that are moving around in mobile networks. The Edge is where the Akamai platform operates, in the last-mile networks that connect end users to the internet. This is very different than where the major cloud companies and the other CDNs place their infrastructure. They usually place their content in a few dozen cloud data centers in the core of the internet. Some call it the Edge, but it's really more in the core, not especially close to end users, and often on the wrong side of congested peering points.
Enterprises and hosting companies also usually locate their data centers in the core of the internet, which is often far from most users and even many of their own employees. In the old days, most companies put all their web infrastructure in a small number of these private or hosted data centers, which were connected to the internet with transit provided by the large tier one networks. Many also use services like AWS and Azure for storage and compute. This helps lower cost and improve agility, there's still the problem of being close to end users and getting the scale you need for today's digital business. For that, you need a true edge platform.
That's why more and more functionality is moving to the Edge, and why most major enterprises use Akamai's Edge platform to deliver and secure their content. The Edge is important because it's where all the users are, it's where all the devices are, and it's where most of the bandwidth is. It's also where most of the attackers and bots are. It's where your defenses need to be. The Edge will become even more important in the future since it's where 5 G will be. This is why everyone is talking about the Edge today. I think it's important to point out that Akamai has not just been talking about the Edge for over 20 years. We've actually been deployed at the Edge for over 20 years. That's one reason why we can provide better performance, higher reliability, greater scale, lower cost, and stronger security than our competitors.
As I mentioned, Bobby will talk a lot more about our Edge platform later today, and now I want to shift gears and talk about the services that run on our Edge platform. Akamai has two main lines of business today, security and content delivery. I'll give you a brief overview of each. Then we'll go deeper with the leader of each business. I'll start with security. Security accounts for about a third of our revenue today, and it's been growing at a very strong rate. Over the last five years, we've become the market leader in several areas, including DDoS prevention, web app firewalls, and bot management. As Ed will talk about later, we have excellent margins in this business, and as it continues to grow, that should help margins for the company as a whole over time.
The security business also helps our content delivery business in terms of bundling to create offers such as Protect and Perform, which has been very popular with customers. Our security business also has great cost and operational synergies with our CDN business, since we use the same infrastructure for both. As we're doing the delivery, we provide the security at the same time. If you buy Kona Site Defender, which is our web app firewall service, you can get DSA or Ion included in a bundle, since we apply the security at the same time that we accelerate the delivery. It's all part of the same data flow with the end user. It seems strange to people, Akamai security products actually have a side effect of speeding up your site. With most other company security products, the security slows things down.
At Akamai, we've engineered our platform to get speed and security at the same time. Our security offerings are grouped into three main areas. Application security, which consists of app and API protection and fraud prevention. Network security, which consists of infrastructure protection and access control. Professional security services. I'll say a few words about each of these categories, working from left to right. App and API protection is all about keeping a site or app safe from attacks by stopping things like cross-site scripting attacks, Magecart attacks, and SQL injection attacks. These attacks are very common and can be used to corrupt the content on a site, inject malware onto the user's browser, and steal private information. Our products in this area include our market-leading Kona Site Defender solution, Web App Protector, and our new Page Integrity Manager.
Last year, these products accounted for $480 million in revenue, which was up 28% over 2019. We have about 2,000 customers using these products, which means that there's plenty of room for growth. Our goal is to sustain a CAGR of about 18%-22% for these products over the next three- five years. Rick will say more about these products and our goals for future growth in a few minutes. Fraud prevention means keeping user accounts and identities safe from attackers who want to exploit them. As an example, there are many attacks today where user credentials are stolen from one site and then used by bot armies on thousands of other sites to see if the same credentials work on multiple sites. When they do, the credentials are sold to a criminal organization that uses them to take over your account.
In fact, this is so pervasive today that most login attempts at many of our customers' sites aren't human. They're bots trying to break in. Our Bot Manager solution identifies and stops these attacks, and it's a big reason why it's been one of our fastest-growing new products over the last decade. Last year, these products accounted for $176 million in revenue, which was up 36% over 2019. We have about 500 customers using these products today, which means the ARPU is very good. That's because of the large value we're providing to customers. As we look out over the next three-five years, our goal is to sustain a CAGR of about 20%-30% in this area. Altogether, we generated $656 million from our app security products last year, which was up 30% over 2019.
This will be the area of greatest absolute dollar growth in the business over the next few years, our goal is to sustain a CAGR of 20%-25% in this area over the next three-five years. Infrastructure protection includes products that block DDoS attacks, like our market-leading Prolexic solution and our Edge DNS service. Prolexic defends data centers at the IP layer, as opposed to Kona Site Defender, which works at the app layer. Edge DNS makes sure that our customer's DNS is operating normally, even in the face of large-scale attacks. These products generated $191 million last year and were slower growing, at 9% over 2019. Going forward, our goal is to grow this category by about 3%-5% per year. Of course, we might do better than that, as we did in 2020, with the increase in ransom DDoS attacks.
Access control is a much faster-growing area and one of our most exciting opportunities for growth over the longer term. It consists of our enterprise security products such as Enterprise Application Access, Enterprise Threat Protector, and our new multi-factor authentication solution, and the products we sell through carriers, such as Secure Mobile and Secure IoT, that were obtained from the Asavie acquisition. These products keep an enterprise safe from malware attacks and data theft attacks. This is a very exciting area for growth as enterprises move away from traditional security defenses and adopt Zero Trust or SASE solutions in the cloud. Last year, these products accounted for only $50 million in revenue, but we're projecting that to double to about $100 million in 2021. Our goal is to sustain a CAGR of at least 50% in this area over the next three- five years.
Rick will talk a lot more about these products shortly. Altogether, we generated $241 million from our network security products last year, which was up 17% over 2019. Over the next several years, our goal is to sustain a CAGR of 20%-25% in this area, with most all the growth coming from our access control products. Many of our security products come embedded with support services, such as 24 by seven monitoring in our Security Operations Command Center. There's no fee for these embedded services, and so we can't report that revenue separately. It's folded into the product revenue. Also, in many cases, customers buy additional professional services where we more proactively manage their security posture. These services are discreetly billed, and last year they accounted for $164 million in revenue, which was up 20% over 2019.
Going forward, our goal is to grow discretely billed professional security services revenue by about 10%-15% per year. Putting it all together, our goal is to maintain a CAGR of at least 20% for our security revenue over the next three- five years. I want to be clear that while 20% is our goal, we're not changing our guidance of 18%-20% growth for security in 2021. Of course, we'll be working very hard to come in at or above the top end of this guidance range, as we've managed to do in each of the last several years. Of course, growth is not always smooth. Some years may be better than others, especially as we integrate acquisitions.
If we can achieve our 20% growth goal, then that would mean we could exit 2023 with a $2 billion run rate for security, and that we could reach approximately $2.5 billion in security revenue by 2025. Of course, it's hard to predict revenue that far out, but we do have several strong tailwinds helping us. Most importantly, cyberattacks are rapidly increasing in frequency, scale, and sophistication. It seems that everyone's getting hacked. Just when you think it can't possibly get any worse, it does. These are just some of the headlines we've seen recently. It's really mind-boggling when you think about all these attacks. The volume of attacks that we're managing at Akamai has also risen dramatically over the last year. For example, we mitigated over 1,800 DDoS attacks in Q4, and that was 40% more than Q4 of 2019.
Our WAF solutions fired almost 2 billion alerts in Q4, 20% more than in Q3. We saw record-breaking numbers of malicious login attempts last year. Almost 200 billion credential abuse attacks, more than four times as many as in 2019. On December 1st, we saw more than a billion in a single day. With all these attacks, our threat research teams are analyzing 290 terabytes of new attack data every day. Now, when you see numbers like these and all the headlines about breaches, you got to ask, why is this happening? Part of the problem is that we're up against some very large, very smart, and very determined adversaries. Malware is literally everywhere, and the traditional defenses just aren't working anymore. Now it's even worse with the pandemic. Aside from the distraction, most companies have a lot of employees working remotely, which increases the exposure.
Since a lot more business is online, the prize for the attacker is greater than ever. This is leading to a perfect storm of attacks, and it's exposing a fundamental issue in the traditional data center defenses. They're just not good enough anymore. Among other things, there's just no way to provision the capacity you need to defend yourself in a data center. Even the largest cloud data centers can be overwhelmed by the attacks we're seeing. Even if it was physically possible to equip a cloud data center with sufficient capacity, you wouldn't be able to afford the cost. The only way to defend against such attacks is to filter the traffic at the edge of the Internet, where most all the capacity is. This is what Akamai does and why we're uniquely able to defend against the largest attacks.
We stop the attacks at the source, at the edge, long before the malicious traffic ever reaches the target. Of course, there's a lot of competition out there looking to grab a share of the security pie. Aside from the large traditional security vendors, there are literally hundreds, if not thousands of security startups and a wide variety of security service providers. Most CDNs now claim to have some security capabilities, even though most are very rudimentary or offered through partners, which often doesn't produce a good end result for the customer. Akamai is performing very well against the competition, because we have the best scale of intelligence. We have the most complete suite of defenses, all on one integrated control plane.
We're constantly innovating and evolving our defensive capabilities to respond to the latest threats. We have a global team with over 1,000 security experts, and we're recognized as a leader by the major analyst firms. For example, Gartner recognizes Akamai as the leader for web application firewalls for the fourth consecutive year. Forrester also names Akamai as a leader in this space. They've also named us a leader in bot management. We're a leader in zero trust for enterprise security. In a few minutes, Rick will talk a lot more about our very exciting future in security. Now I want to switch topics and talk about our CDN business. Our CDN business accounts for about 2/3 of our revenue today and has been growing in the low single digits over the last several years.
Akamai created the CDN industry as we all know it today, and we're the market leader by a wide margin. That's due in part to our massive global scale, superior performance, high reliability, cost competitiveness, and natural integration with our market-leading security products. As Ed will talk about later, we generate a lot of cash from this business, which we're using to invest in exciting new product areas such as IoT, 5 G, and the next generation of edge computing. Our CDN offerings are grouped into three main areas, edge delivery, edge applications, and professional services. The large majority of our CDN revenue is from our edge delivery solutions. This includes video streaming, software downloads, and accelerating sites and apps with our market-leading Ion solution. As we discussed in our recent earnings calls, we expect this business to continue to grow in the low single digits over the next several years.
2021 may be flat or slightly down, depending on the impact of the pandemic on customers in some key verticals like travel and hospitality, and the tough year-over-year compare in our media vertical, which saw explosive traffic growth in 2020. The edge delivery space is highly competitive, per unit pricing typically decreases over time as volumes rise. The good news is that our delivery revenue is highly diversified. We have some very large customers, only 10 that account for more than 1% of our overall revenue. Those 10, put all together, account for less than 20% of our overall revenue. Edge applications accounts for a small portion of our CDN revenue today, is growing very rapidly, primarily driven by our latest edge computing offers.
If you assume a broad definition of edge computing, as many do in the community, then the revenue we derive from edge computing at Akamai far exceeds what we're showing in this bucket. The reason is that we don't charge customers separately for many situations where they're developing and deploying apps on our edge platform. It's billed as part of their Ion or DSA contract, and so it's counted in the edge delivery category. What we're showing in this bucket is just the revenue from edge computing services where it's billed separately, which is the case for some apps that run on the edge for things like image management, as well as apps developed by our customers and third parties using our latest JavaScript technology. As you can see, the use of these services is growing rapidly, up 23% over 2019.
Our goal is to sustain a CAGR of at least 30% in this area over the next several years. As with security, many of our CDN products come with embedded support services. There's no fee for these embedded services, we can't report that revenue separately. It's counted in the product revenue. What we're reporting in the services box shown here is discretely billed professional services, which accounted for $269 million in revenue last year. Going forward, we expect modest growth for these services as we enable our customers and partners to take on more responsibility for managing their sites and apps themselves. Putting it all together, our goal is to grow the CDN business at a CAGR of about 2%-5% over the next three- five years.
As I mentioned earlier, we expect the growth to be a little short of this range in 2021, but as the edge applications area grows, its much higher growth rate should help improve the growth rate of the CDN business as a whole. Of course, the growth of the CDN business will depend a lot on traffic growth, which has been very strong over the last year. This is a graph of our aggregate traffic over the last three years. The spikes correspond to major events like a movie release, a sporting event, or a major software release. Aside from the spikes, which are a common occurrence at Akamai, you can see the impact of COVID-19 on our overall traffic beginning last March. The big question is how much traffic there will be in the future. For this year, we are projecting more normal levels of traffic growth.
Over the long term, the traffic levels could be dramatically higher than they are today, as more video is watched online and at higher quality levels. You can imagine a world where two and a half billion people watch video every evening during prime time. Already, two and a half people have watched a video online. You can also imagine them watching that video at a reasonable quality at about 10 MB per second for each stream. We already have many customers who are broadcasting at a higher rate, and today, 10 MB per second is the desired goal for anything that's going to be watched on a form factor like a TV. Now multiplying two and a half billion people by 10 MB a second, that gives you 25,000 TB per second. 25,000 TB per second is a lot of traffic.
It's orders of magnitude beyond what can be served from traditional cloud data centers. There is enough capacity to serve this volume of traffic at the edge. In many cities, the broadband connection into the typical home or office has tens or even hundreds of megabits per second of capacity, more than enough to download a large software file or to watch a high-quality video online, and it'll get even better with 5 G. Overall, the capacity at the edge is about 100x the capacity in the core. As more media moves online, serving it from the edge will become even more important. Being at the edge close to end users also provides better quality and reliability, and this is why so many of the world's largest media companies rely on Akamai to deliver so much of their traffic.
Of course, it's not all about scale and traffic. Performance also really matters to companies not in the media vertical. Performance impacts their brand, search rankings, conversion rates, revenue, and ultimately their customer satisfaction and loyalty. Simply put, performance is vital to just about everything you do in online business. With COVID-19 causing more businesses to move more of their operations online, performance matters now more than ever before. That's why so many of the world's major enterprises turn to Akamai to accelerate their sites and apps. Because our servers are at the edge, they're closer to the end user, which means that the latency will be less. If you're serving content from the edge, you can avoid the congested peering points in the core of the Internet. This is pretty obvious for static or cacheable content.
What's a little more surprising is that by having our servers close to the end user, we can also deliver dynamic or non-cacheable content faster. That's because we've developed our own routing and communication protocols to find fast paths back to the origin where things like your bank balance or account information is stored, so it can be retrieved quickly. Also, if there's a problem with a peering point getting congested or a route being dropped on the Internet, we'll find another path back to the data center automatically, which greatly increases reliability. If your origin is having problems, we'll automatically serve up alternate content or direct users elsewhere for you. We also prefetch the content on any given page for each client, so it's there just before it gets requested.
This kind of just-in-time caching makes it feel like the user is right next to the data center in terms of speed, even when they're very far away. By having apps run on the edge, speed can be improved even more, and you can offload the data center. This is sometimes called serverless computing, and it's especially important for apps that need to scale instantly and on demand. Akamai was the first company to offer edge computing services, and we've been doing it for over 20 years. We began in 1999 with advanced metadata, which is an XML-based language used to enable a variety of capabilities on our edge platform. Things like access revocation, ad insertion, throttling, and visitor prioritization.
In 2001, we introduced Edge Side Includes, which allows our customers to deploy logic on our edge servers to assemble web pages dynamically based on information about the user, such as their location, device type, and cookie information. In 2002, we released EdgeJava, which allowed customers to deploy their Java code on our edge platform. IBM created a version of WebSphere that ran on our edge platform in 2004. This was then used to deploy a wide variety of apps on our edge, things like car configurators, online contests, voting, user registration, and managing shopping carts. As it turned out, we were ahead of the market with edge computing. AWS hadn't even been started yet. Most customers weren't ready to migrate their apps to our platform. They could do it, but it wasn't as easy as it is today.
We decided to build apps ourselves for common tasks like video management, visitor prioritization, and application load balancing. This worked well because the customer didn't have to do any programming or porting. With the advent and widespread adoption of cloud services like AWS and Azure, the market is now much more capable of writing apps and microservices that are designed to work in a cloud environment. In 2019, we created EdgeWorkers so that customers can easily port their JavaScript apps to our Edge platform. Customers were already using it for things like API orchestration, personalization, search engine optimization, compliance, and virtual waiting rooms. Third parties are also beginning to write apps using EdgeWorkers, which creates a very interesting future potential of an app marketplace running on our Edge platform.
Later today, Adam will talk more about this and mention some of the early examples of third-party apps running on our platform using EdgeWorkers. By the way, I should mention that while some people refer to edge computing as serverless computing, it's really not serverless. It's just that the apps run on our platform instead of on the customer's infrastructure. From the customer's perspective, they don't need to worry about any server infrastructure, and they can scale up instantly on Akamai. There are many companies that offer some form of CDN services today. Akamai is the market leader because of our superior performance and reliability, our unparalleled global reach and scale on demand, excellent video quality and fast software downloads, our expert services and support, and our ability to provide the complete package of performance and security services.
When it comes to edge computing, we offer native JavaScript support and very fast startup and response times. These are some of the reasons why our market share is so strong, as recognized by this report from IDC. The size of the dot correlates to how much market share each company has. As you can see by the size of the big orange dot, Akamai is doing very well. Being in the top right corner of the chart means that we rank number one when it comes to both strategy and capabilities.
Of course, the real test is customer adoption, and it's great to see Akamai is used by so many of the world's leading brands to deliver, accelerate, and or secure their applications, including over half of all Fortune 500 companies, over 800 e-retailers, including 17 of the top 20, over 260 hotel and travel companies worldwide, including 41 of the top 50, over 300 banks worldwide, including 15 of the top 20 globally and all top 29 in the U.S. Over 200 government agencies worldwide, over 275 telcos, carriers, and ISPs worldwide, including all top 50 carriers, over 225 game publishers worldwide, including 23 of the top 25. Over 200 OTT video companies, including nine of the top 10 in the U.S. and 28 of the top 30 in EMEA. In summary, I think Akamai is very well-positioned for future success. The fundamentals of our business are strong.
We're committed to maintaining high margins while driving growth, and we're focused on innovation and operational excellence. We have excellent prospects for future growth with our innovative and rapidly evolving portfolio of integrated products. We're the market leader in CDN with an exciting and fast-growing Edge applications business. We're also the market leader in the fast-growing application security space, with especially strong growth in the area of network access security. Our market position is very strong with our unparalleled global Edge platform, our premier customer base, our strong profitability, which we can use to fuel the investments needed for future growth, our industry-leading technical expertise, and our natural integration of security with delivery and performance products. We also have a seasoned management team with a well-established track record of success, having created two industry-leading billion-dollar businesses, CDN and web security.
This team has also proven their operational excellence by growing revenue at a five year CAGR of 8% and non-GAAP EPS at 16%. We also have several powerful secular tailwinds helping us. OTT is still in the early days with much more traffic coming online. The move to online for work and commerce is here to stay. Cyber attacks will continue to grow in scale and sophistication, and 5 G will provide a quantum leap in device connectivity, truly enabling IoT and creating unprecedented demand for edge computing. In short, the Akamai Edge will be needed in the future more than ever before. That concludes my presentation. Thanks very much for listening, and I hope you enjoy the rest of the talks that we prepared for you. I look forward to taking your questions at the end.
Next, you'll be hearing from Rick McConnell, who is our president and the GM of our Security Technology Group.
Thanks, Tom. Hello, everyone. I am delighted to be with you today to talk about Akamai's security business. I've actually been responsible for our security portfolio since I joined Akamai back in 2011, when security revenue was in the low single-digit millions of dollars. Having now surpassed $1 billion in 2020, growing at 25%, I couldn't be more proud of what our team has accomplished. Even more importantly, I couldn't be more excited about what lies ahead. Now let's talk about our Security Technology Group, and Adam will follow in a bit with our Edge Technology Group. In security, we have now combined all of the security components of overall Akamai Security into one. Essentially bringing together our web enterprise and carrier security products under one umbrella within this group.
Our goal, quite simply, is to grow that billion-dollar business to $2.5 billion, so well more than doubling it over the years to come by 2025. Let's talk a little bit about the framework that Tom covered earlier, which consists of application security, network security, and security services. At the high level, our three- five-year goal is to grow it at 20%+ from the billion-dollar starting point in 2020. As Tom indicated, this does not change our 2021 guidance of 18%-20% growth, but as we look ahead with a combination of organic and inorganic growth, we believe the 20%+ three- five-year goal is achievable.
Our security services bucket consists of managed and professional services billed discreetly to customers versus implementation and other services that are embedded in the product sales buckets. I'm going to focus, though, on the two major components of the business and our portfolio. First, let's look at market size, which is substantial. Within application security, we believe the total addressable market is approximately $12 billion. Within network security, an even higher $17 billion. This gives us a $29 billion overall security TAM for Akamai across these two businesses, which obviously affords us substantial runway as we look ahead. We look at the competitive landscape across many dimensions and against numerous categories and competitors, including hardware providers, CDNs, cloud providers, and security point solutions as you see on the left.
Whether in scale, real-time threat data, security services capabilities, and other areas, there is no competitor that has the breadth or depth of capabilities that we have in our target markets. That is further supported by broad analyst validation, some of which Tom covered earlier with numerous leadership indicators. As we look at some of these comments, which as I say, Tom covered, this includes such designations as leadership in Gartner's Magic Quadrant for Web Application Firewalls multiple years running. Numerous Forrester Wave leadership positions and support on our existing and new products like Page Integrity Manager from 451 Research. I love the quote at the bottom, though, from IDC, which says that Akamai has had the strongest and broadest edge security offering for quite some time, and we believe that continues to hold true as a core differentiator of Akamai's overall product portfolio.
There are several overall growth drivers for the business. Let's start with new customers. New customers at Akamai really are being driven in their deployment of solutions by security. That is what keeps them online. It protects their business. The fact that they can get security at the best performance at the most significant scale drives the differentiation for Akamai in new customer acquisition. Installed base is another area. There are actually two dimensions to the installed base deployments and sales. The first is an add-on sale of security to our existing CDN customers. CDN customer might choose, for example, to deploy Kona Site Defender, our web application firewall solution, as their initial security deployment. We have an opportunity to add on sales of security solutions, perhaps adding on Prolexic or Bot Manager or upcoming Account Protector product, Page Integrity Manager, and others.
Multiple product deployments gives us a substantial opportunity for incremental growth. We can expand to adjacent markets as well. We've done that with Enterprise Access, Page Integrity, manual fraud, multi-factor authentication, and more to come. These adjacent markets provide us incremental growth opportunity. Finally, I'd simply note channel and geo expansion. Last year, we acquired Exceda, for example, to expand in Latin America, and that's gone extremely well in bringing our security products to yet another geo more aggressively than we were able to do before. Tom showed our overall customer leadership position previously, but this holds true as well for our security customers. As you can see, we have very strong positions in e-commerce, high tech, financial services companies, airlines, manufacturers, state and local government, OTT providers, and others.
These companies typically view Akamai importantly as mission critical in enabling and protecting their online presence. I now want to turn our attention to our application security business. As Tom mentioned, this business delivered $656 million in revenue for us in 2020 with 30% year-over-year growth. Our three- five-year growth goal is 20%-25%. Breaking this down, this consists first of App & API Protector, which is the flagship area of our security offerings. This business delivered $480 million in 2020 revenue, growing at 28% year-over-year. The growth for this business is expected to be reasonably stable over the three- five-year horizon in the 18%-22% CAGR range. For fraud prevention, this has been a phenomenal area of growth for us, delivering $176 million in 2020 revenue at a 36% year-over-year growth rate.
As the Bot Manager revenue has grown substantially, we expect that growth rate in this business to come down in 2021, but then pick back up in 2022 to 2025 with the delivery of new products such as Account Protector and potentially inorganic growth, resulting in an expected 20%-30% CAGR over the three- five-year horizon. We spend a lot of time thinking about penetration of our security products into our installed base. The chart on the right shows our penetration of our main application security products into Akamai's top 500 customers. While our web application firewall products have nearly 80% penetration into this customer base due to their time in the market, we still are under 50% deployed with Bot Manager, and we have really just started with Page Integrity Manager as of April of last year.
There is a lot of white space still to come. Additionally, if you look at Akamai's overall customer base beyond the top 500, we are still only 41% penetrated with at least one application security product. We have ample opportunity for growth even with current customers. Finally, it is worth noting that we see much higher switching barriers for customers who have deployed at least one security product. Their typical churn rate is already extremely low, in the low single digits for Akamai overall. With one security product deployed, that churn rate gets cut in half, and with more than one security product deployed, the churn rate falls by more than 75%. All of these churn rates, as I said, are extremely low. Let me now go on and talk about the product portfolio in a bit more detail. First, beginning with app and API protection.
Web Application Protector is our mid-market-oriented product with automated delivery of capabilities for web application firewall. Kona Site Defender is our long-term flagship solution for web application firewall, provided to customers, and the combination of these two solutions, as I said earlier, have very substantial deployment within our installed base. As you can see some of the stats at the bottom, we have enormous capabilities in terms of scale. Almost 2 billion attacks mitigated in the fourth quarter alone, and approximately just shy of 300 TB of threat data evaluated each day in Q4 of 2020. We're looking to expand with built-in bot management, API discovery, and risk profiling to expand these product lines. Page Integrity Manager we launched back in April of last year, April of 2020, and it is designed to address formjacking and third-party scripting, which can be malicious.
The result of that is that we can enable much, much greater protection for customers where potentially malicious capabilities have been injected into these third-party scripts. It has absolutely exceeded our expectations in the market thus far since April of last year over its first year, and you can see we're already protecting monthly nearly 3 billion pages. We have lots of investment and development yet to come in this evolving space, and we're delighted with the pickup from customers thus far. Now let's turn to our fraud prevention area, where our flagship solution is Bot Manager. Bot Manager enables advanced bot detection to spot and avert the most evasive threats. Now, Bot Manager began just as a way to categorize bots. Some bots are great, like SEO, for example. Other bots, not so much when they're trying to steal credentials.
What we did was we expanded that solution to include credential stuffing with Bot Manager Premier. That then further fueled the revenue growth opportunity in this space and in fact, our overall revenue growth. As you can see, over 70 billion attacks mitigated during the fourth quarter of 2020, and substantial investments here as well in a rapidly growing market opportunity. Things, for example, like bot risk scoring. I also really like our opportunity here in managed bot services. Over in the right column of that framework slide, we talked about managed services and professional services briefly. Most of those are delivered 24/7 in monitoring of our Kona Site Defender services with Managed Kona. We believe we can offer managed services associated with Managed Page Integrity Manager as well as Managed Bot Manager as incremental solutions for those security offerings for security services.
Account Protector is a manual fraud solution. Whereas Bot Manager is oriented at automated bot protection, Account Protector is oriented at manual fraud. This is an obvious next evolution of our overall fraud prevention strategy. As we have said before, we believe that we've only just begun in the opportunity ahead for Akamai in this important area of fraud prevention overall. In terms of case studies, I don't want to spend a lot of time on these, but suffice it to say that we have very, very strong customer capabilities who are delighted with these services, who provide us this feedback frequently. In the case of financial services on the left, for example, a huge credential abuse attack, over 11 million credential abuse attempts. We solved it for them with a combination of Kona as well as Bot Manager.
On the right side, we had an e-commerce customer that was getting targeted by Magecart attacks, which is formjacking, and the result of that was that Page Integrity Manager got deployed to protect that customer and improve their own code base looking forward to ensure better protections are in place looking ahead. Let's turn our attention to the second area of business for us, network security. In network security, that business delivered $241 million in revenue in 2020 at a 17% year-over-year growth rate. Our three-five -year growth goal here, as with application security, is 20%-25%. The first component of network security is infrastructure protection, which includes our DDoS and DNS solutions. This business delivered $191 million in 2020 revenue, growing at 9% year-over-year.
The growth rate for this business is expected to be stable over a three-five-year horizon in the 3%-5% CAGR range. As we saw in 2020, however, this growth can be higher if we see an increased occurrence of attacks in the market. The second, and our most rapidly growing security segment is access control. This business accounted for $50 million of 2020 revenue, which grew 62% year-over-year. But as Tom mentioned, we are anticipating this business to grow by more than 100% to over $100 million in 2021. From there, as the business size increases, we of course expect growth to moderate, but we are still anticipating a 50%+ CAGR over a three-five-year period. This is another area in which we are looking at inorganic growth to supplement our internal investments.
As with application security, there is substantial white space for our network security within Akamai's installed customer base. For our infrastructure protection products, Edge DNS has 50% penetration into our top 500 customers, while Prolexic is 25% deployed. Our enterprise access solutions have even more runway for growth. We also have a lot of opportunity with our secure business and secure mobile products with the top carriers. Again, we see higher switching barriers and higher opportunity for wallet share expansion with security products deployed. Now let's take a look at some of the individual products within network security. First and foremost is Prolexic, which is a layer three, layer four DDoS protection service to assist our customers in protecting the data center. While Edge DNS enables another layer of DNS security for large DNS attacks of our customers that happen through authoritative DNS.
We have a huge number of these DDoS attacks mitigated year-over-year, 40% growth, in that statistic. There is a lot of opportunity for expansion in each of these areas that we're making, from intrusion detection and prevention, network detection and response, IDS or NDR as they are called with Prolexic, and the areas of Edge DNS we're combining into that solution overall our traffic management capabilities to provide a more integrated solution there. As we move to our enterprise access capabilities, I couldn't be more excited about this part of our portfolio. The overall network topology is absolutely changing right before our eyes. At Akamai, our new headquarters building doesn't use VPN. We simply get online, and our devices and the access credentials that we provide essentially control how we access what we access and how we get on the network. Our enterprise access capability provides this access.
Enterprise Threat Protector then looks for malware through a recursive DNS solution that enables us to evaluate much better how we can provide access and what we provide access to out of a wide array of devices, whether they be laptops or cell phones or IoT devices or others. This is changing the network topology, and it is fair to say that the walled garden of old through VPNs is simply quaint. It is not going to work in the future because it is too easy through phishing or other solutions to infuse the network topology. A change in this network topology toward things like Zero Trust Network Access makes a ton of sense as we look forward as that network topology changes. As you can see, we're already looking at analyzing more than 3 trillion DNS requests every day. A huge number.
We're blocking 70 million or more malicious DNS requests each day as well in these areas. Again, we have a ton of investment opportunity as we look ahead, ranging from data loss prevention with DLP to other enterprise integrations to simplifying our overall solution set for deployment into enterprise environments to device posture, signaling and other integrations. As we look in our carrier solutions with secure business and secure mobile, we also see substantial opportunity for incremental growth. Secure business essentially provides similar recursive DNS solutions to provide protection to small businesses, to themselves as well as their customers.
Secure mobile came out of our Asavie acquisition, which we just completed last quarter. That provides us a capability to provide essentially IoT and other device management on the network to provide that next layer of security for access of these devices to ensure no malicious activity is occurring on those devices or infusing the network. We've only just begun in many ways in these areas, and it's resulting in substantial growth in our access control area, which I talked about earlier. As we look at the case studies here in network security, we see, first of all, in online banking, targeted DDoS solutions and emergency onboarding. This came out of the extortion and ransom attacks of Q3 and Q4 of 2020. We deployed dozens of customers in this way in an emergency setting to ensure their protection of their network amidst these extortion threats.
In gaming, we see that with The Pokemon Company, for example, pursuing zero trust, they decided that they wanted to change and morph their overall network topology. We helped them do this with EAA, Enterprise Application Access, as well as Enterprise Threat Protector. Finally, we have numerous carriers, one of which is Comcast, that deployed secure business and had enormous take-up in our overall customer base, or their customer base, I should say, through our offerings of the secure business opportunity. Bringing it to a close, as I said earlier, we have an exceptional opportunity at Akamai to grow our security business. We are providing extraordinary value in protecting our customers online. We have the most extensive real-time threat data, an extremely broad portfolio, an extraordinary and sizable team of security experts, and the benefit of our Edge platform and its enormous scale and performance.
We also have an ability to grow along many dimensions, from new customer acquisition to new product lines from both organic and inorganic expansion to extended channel and international presence. I'm excited about 2021, and I'm enthusiastically looking forward to continuing to see very strong growth in security for Akamai in the years to come. With that, I'd now like to turn it over to Adam Karon, my colleague and Akamai's Chief Operating Officer, as well as the GM of our Edge Technology Group. Adam?
Thanks, Rick. Appreciate it. I'm really looking forward to walking everybody through an overview of our Edge Technology Group and answering all your questions at the end of the session. With that, let's get started and talk about the organization itself. The Edge Technology Group brings together our delivery and compute products, along with all the teams that build and operate Akamai's Intelligent Edge platform. We have a vision to power the edge of digital innovation, enabling immersive experiences with performance, scale, reliability, and simplicity across any application architecture, all with the goal of ensuring that the Akamai platform remains the unparalleled market leader for scale, performance, reliability, ease of use, and of course, cost. Now, we break our CDN products into three main groups. First, edge delivery, edge application second, and professional services third.
As you can see, at $1.7 billion in 2020, a vast majority of our CDN revenue comes from our edge delivery solutions. This includes video streaming, software downloads, accelerating sites and apps with products like Ion and Adaptive Media Delivery. As we've talked about in our recent calls, we saw this business grow at around 3% in 2020, and we expect it to continue to grow in the low single digits in the coming years. That said, edge delivery for 2021 might be slightly down, just depending on the impact of the pandemic in verticals like travel and hospitality. While we hope to see recovery in these verticals quickly, we expect that it'll really be in the back half of 2022 or into 2023 before you really see something.
Let's not forget that we have a really tough compare in 2021 from 2020 in terms of our media vertical, because COVID drove tremendous traffic growth over the last year. As Tom mentioned, the industry is highly competitive, and it does have the characteristic that as volume or traffic increases, you do see unit pricing coming down. As he said, it's a very large customer base, and it's very highly diversified. The next bucket, we have our edge applications area, which I'm really excited about. Today, it accounts for around $151 million, but it's very fast-growing, and it's heavily driven by our latest discreetly billed edge computing products.
Examples in this bucket are apps that run on the edge for things like image management, visitor prioritization, contests, voting, as well as apps developed by our customers and third parties using our latest EdgeWorkers product, leveraging JavaScript. This is just the beginning of growth in this area. While it was up 23% over 2019, our real goal is to sustain that CAGR of 30% in this area over the next several years. As Rick mentioned in security, many of the CDN products come with embedded people services, and there's no separate fee for these services, like 24 by seven support. We don't report that revenue separately. We do count here is discreetly billed professional services, which accounted for $269 million in revenue last year, and 8% year-over-year growth.
Going forward, though, we expect the services for this part of the company to grow slightly modestly here as customers and partners take on more responsibility for managing their sites and apps themselves. All right. When you bring it all together, the goal is to grow the CDN business at a CAGR of around 2%-5% over the next three- five years. As the edge application area grows, its much higher growth rate should help improve the growth of the CDN business as a whole. As you can see from the analyst TAMs, $9 billion in each section, we have a lot to go after here. In terms of growth drivers, we have the great fortune to be working in markets that are both growing tremendously and navigating through fundamental shifts in business models as well as technology platforms.
This provides a rich tapestry of opportunity within our media business in both video and gaming segments. The opportunities in video are abundant. We're all well aware that we're in the midst of a massive shift in the video market. Consumers have led that shift with changing habits in how, when, and where they watch. Akamai media products have been, and always will be, right at the heart of that shift. While video is all about navigating a transition to a new future that COVID has only served to accelerate, we see our gaming customers pushing the envelope every chance they get. There's a number of trends driving the gaming industry. The first is free-to-play games, which are now among the most profitable games in history. The games allow everybody to get their hands on the game with no barrier to entry.
It's free, it drives significant download traffic. The same games require exceptional player experiences. The slightest issue can drive players away. The fact that the games are free means early players who get frustrated just leave. They have no sunk cost mindset, you need to keep it good to keep them loyal. Another is multiplayer gaming, it's a key driver for long-term success. A living game as a service keeps players engaged over the long haul. It drives the need for robust security and reliability to keep games running 24 by seven. The last is the transition to an all digital future. As game delivery continues its migration from being delivered via disc to being delivered digitally, it encourages a culture of constant updates, connected players have come to expect new content and constant improvements.
We see that in the quarterly season updates from some of our favorite games. This means more and more data to devices, more immersive experiences, and more need for high performance internet routing for complex gameplay. Our platform for distribution acceleration and advanced gameplay is a critical part of gaming today and in the future. The last two I'll talk about more in depth towards the end of my discussion, but suffice it to say that both Edge Compute and 5 G will provide tailwinds to Akamai's business by driving more and more compute to migrate from the centralized clouds to Akamai. Our Edge network will be able to take full advantage of the promise of 5 G. I'm going to save both of those for a little bit later. Next, I wanted to talk about the competitive position in the market.
There's a lot on this slide, and I won't drain the slide. I think it provides a good reminder of the power of a platform like Akamai's. Having a platform where security, serverless compute, and delivery can all be delivered on the same network anywhere in the globe, very close to an end user, provides differentiation that none of the competitor segments can offer. I think it's important to stress that underneath these features is our robust relationship with carriers all over the globe, providing localized access for our products where others just can't compete. The power of a single platform is a major advantage to our continued success. When you think about independent validation of that unique market position, you can look no further than IDC's view of the marketscape that Tom showed earlier.
Akamai leads the space where the others are really just playing catch up. How about traffic growth? Tom showed a view earlier, I thought I'd show it just a little differently here. I like the view with the events on it. Tom talked about a day when we'd see 25,000 TB. When you think about how far we've come in just a few years, it's really mind-blowing. I vividly remember thinking that the peaks of 1.3 TB for the royal wedding were huge. Today, we think about those types of peaks as being very tiny when we're driving 180 TB or more on the platform.
That 25,000 TB day is something we used to think is impossibly far away, but with video and gaming consumption trends heading in the direction they're heading, you can really start to see that reality coming closer. Customers across the world trust Akamai to deliver the most important video and games to their end users everywhere. Whether it's a critical live sporting event, an on-demand launch of a new movie, a brand new game release, maybe urgent news release to the public, Akamai is the choice of major brands all over the world. That brings me back to a reminder of our customer penetration with our CDN products. Now, Tom went through this slide earlier as well, so I really won't do that to you again.
I won't read the slide, but I do want to stress that we are highly utilized in all verticals and by the top companies and brands in all of those verticals. Now, let me go a little bit deeper into the CDN offerings that give us that strong market position, starting with our Edge delivery products. Within our Edge delivery bucket, we have products which are focused on the unique needs of video streaming that include things like optimization for use cases such as looking ahead and getting the next section of video before the playback device even asks for it to provide the best user experience possible.
We also offer managed CDN solutions and virtualized or licensed CDN solutions as dedicated delivery products, which we sell to carriers that take full advantage of the features in our video delivery products and make it available in ways that carriers need to implement within their own networks. Our download delivery product is designed for the gaming and software download use case, and it includes specialized capabilities like indicating whether a download is in the foreground or background. Maybe it adjusts features, even performance, based on specific needs of the customer. We also offer supporting products that will differentiate us in performance, reach, and scale. As an example, NetStorage is a globally distributed storage platform that includes sophisticated rules for replicating content for availability and performance.
Our media services live product ingests live stream content and enables very unique features such as really low latency and streaming to broad geographic locations at extremely high quality. We offer visibility through our analytics products. Our media analytics provides customers real-time data on viewer activity and couple that with our BOSS and BOCC, and it goes even further to connect all the pieces of the video workflow into a single integrated view, combined with our leading professional services monitoring and addressing any issues that just might arise. Major media outlets and gaming publishers you see on this slide, everybody from AT&T to Fox to Sky, maybe Pokémon like Rick talked about, or Roblox, maybe Fortnite. They all depend on our media delivery solutions to provide performance, scale, security, and reliability for all of their most critical online needs.
Let me pivot into a deep dive into the details of our application performance products like Ion and DSA. To start, these application delivery solutions inherit all of those performance, reliability, security, and scale of our Intelligent Edge platform. When you add on top of that the power of real user data powered solution like our mPulse product, it not only enables real user-based analytics, empowering our customers to make business decisions, but in addition, it also provides the unique value of using this real user data to power low touch, automated performance optimizations. These optimizations include things like automatic push, pre-connect, and preload of critical resources, automatic Brotli resource compression, reducing the page weight. Visibility and protection against non-performant third-party scripts with script management capability.
Now with analytics in place and the performance optimized experience, our solution additionally offers the ability for customers to do what if analysis with testing using our highly scalable global load testing platform called CloudTest. This enables customers to safely simulate large events on their site, ensuring they're always prepared. Now, we're constantly investing in our platform, ensuring optimization of key metrics such as time to first byte and offload are perfect. Just in the last year, this investment has seen results of reducing origin traffic of up to 65% for our web application customers, improving the eviction age threefold and improving turnaround time as part of time to first byte by 25% at the 99th percentile.
We're also investing in the latest technologies such as H3 and gRPC, as well as automated optimization, including features to augment our already existing adaptive acceleration solution with the introduction of Brotli on the fly, providing for real-time Brotli compression across dynamic and static content to minimize page weight and ultimately increase performance. We're always focused on customer pain points and use cases. Geo Delivery is one such example of a use case-based solution that allows customers to define traffic restrictions by geography in support of things like, tax implications or maybe privacy compliance. We provide these solutions for many major brands, and I'm going to touch on two specific examples that really had to do with the changes in their industries due to COVID. The first one surrounds dynamics in e-commerce.
Retailers all across the globe saw a surge in traffic as the majority of their businesses shifted online. Major retailers like Alibaba saw tremendous growth. During Alibaba's Singles' Day event, which is similar to our Black Friday day events across multiple retailers, Alibaba's traffic through Akamai increased over 173% over the year before, delivering a massive peak for an e-commerce site of 457 gigs per second of traffic during just the first 30 minutes of the event. It enabled Alibaba to seamlessly support 26% increase in sales over 2019. While we did this, we did this for many other customers, all simultaneously, either Chinese or anywhere in the globe, illustrating the scale and power of the Akamai platform. The second example revolves around the restaurant dynamic in 2020, which also changed quite dramatically due to COVID.
Many restaurants and restaurant chains were experimenting with online ordering, but really a few in the beginning saw much revenue from it. All of their revenues really shifted to online ordering and curbside pickup, they were forced to not only dramatically improve the scale of their sites and also their reliability, but they had to add a lot of new features and capabilities just to keep up. One major North American retail restaurant chain floundered in the early days, right? They tried using a competitive CDN to make their site work during the pandemic. They brought in Akamai worked with them to improve their visibility through our performance analytic tool, mPulse. mPulse helped them identify problematic areas for scale and performance. We quickly deployed changes to improving scale and availability with our award-winning Ion solution.
Let me go deeper into the second area of our CDN offerings, our edge applications area. Akamai is the only company that has been building best-in-class edge compute applications for 20 years. We've created the building blocks needed to build the most scalable, performant, reliable, secure, and compliant edge applications. Customers will now have access to the same building blocks we've used internally to build their own edge-native applications. To be specific, Tom showed this slide. He covered all 20 years of our innovation in edge computing. I'll only be focused on the edge computing solutions that aren't grayed out, the stuff you see to the right here on this slide, it's the items that Akamai bills discreetly for, not compute that might be included in our delivery products.
This is the specific area where we intend to drive that 30% CAGR that I spoke about earlier. All right. EdgeWorkers is the first product. It's a globally distributed serverless platform that enables our customers to write and deploy code without having to worry about scale, redundancy, and any other overhead. We've deployed Chrome V8 engines to our edge servers, which effectively turns our, well, the world's most distributed CDN footprint into the world's most distributed serverless platform, where every single one of our edge servers becomes a compute node. The code executes at the closest node to the end user, enabling our customers to deliver a performant, scalable, reliable, and consistently, really just awesome user experience. Our customers write code in JavaScript. They just push it out to our network. That's it. They're done.
We purposely focused on JavaScript to start, since it's a language that our customers already know, lowering the friction for them to get started. By being integrated right into the CDN flow, our customers will have insight into just every request, enabling them to build customized experiences for their end users. We'll be continuing to make exciting enhancements to EdgeWorkers, including surfacing insights from other products so that the customer can use our EdgeWorkers as the glue which connects all Akamai products together to quickly solve whatever their custom use case might be. On the right, we have our EdgeKV. It's a globally distributed key value store. By being tightly integrated into our CDN, our customers are able to solve use cases that require cache feed reads and local data processing. This enables our customers to implement edge native workloads where state is required.
When you think about state, think about something like reading and writing a real-time inventory catalog where inventory needs to be up to date right now. Edge KV is tightly integrated into Akamai's Edge, which means our customers can deliver user experiences that require those fast reads and writes with extremely high availability. We'll be continuing to enhance the Edge KV to provide our customers with the ability to adhere to data regulations across the globe. Next, we have first-party applications, like Image and Video Manager or maybe our Cloudlets suite of products, which included things like Visitor Prioritization that provide our customers with an easy-to-use waiting room, or maybe our audience segmentation app, which helps our customers perform A/B testing at the edge. These products have seen tremendous adoption over the years.
In fact, our Image & Video Manager has seen nearly 100% year-over-year growth in the number of variant images that we created for customers, with over a trillion images created in 2020 alone. Now we're creating over 900 million optimized images and videos every day, and we hit a peak of nearly 1.5 billion created on the Singles' Day that I spoke about earlier with Alibaba. Finally, we have a focus area around IoT on the right. Our objective is to empower real-time use cases, enabling our customers to transport data to and from millions of IoT devices around the world by providing the world's most distributed message broker, which supports IoT protocols like MQTT. IoT Edge Connect is a real-time messaging platform that enables near real-time IoT use cases at scale.
Our IoT OTA Updates enables our automotive customers to deliver updated software in a fast, scalable, and secure way. In fact, our customers used our IoT OTA Updates to deliver updates to more than 15 million connected cars, including mine, last year. All right. Now, on a lot of these slides, I've been saying that we have the world's most distributed serverless platform, what does that actually mean? What does it mean by having the most distributed platform? Well, it means we have a platform that can execute code across hundreds of thousands of servers in over 135 countries, over 4,000 locations all around the world. It enables our customers to deliver performant, scalable, and reliable experiences to over 1.3 billion devices that we connect every single day. Something no other platform can do.
We provide scale and reliability through a single global network that protects against network congestion and ensures reliability and scale with no management overhead. We have distributed compute through a globally distributed edge, where nodes have the ability to execute code closest to the device and the end user, which will be critical in 5G, resulting in decreased latency, decreased bandwidth utilization, and increased origin offload. We can process data where it is created by having business logic to process that data right at the edge, close to the device that's using it, to support latency-sensitive workloads and compliance regulations. I thought I'd cover just a few real-world examples here as well. The first one, we worked with a global media measurement and analytics company that needed to manage compliance with an increasing number of regional data protection and privacy regulations all over the globe.
You've heard of them, GDPr, CCPA, APPI, and there's a whole slew of other ones. This particular one required ensuring that tracking data for their publishers' users was consensual. You've seen that before, where you have to consent to being tracked on a website to get access to content. By leveraging compliance tool suites like the IAB's TCF 2.0, which is an industry standard for sending and verifying user consent, the company used our EdgeWorkers solution to build an edge native microservice to determine if a user has allowed tracking data to be collected for their digital experience. If the user, like you or me, consented, the state tracking cookie is deposited, and the user gets a personalized experience on the site. If not, the cookie's discarded, no tracking is applied to the session, and that user gets a generalized experience.
In the second example, in the middle, talk about Fanatics. Query strings added by search engines and social media sites were decreasing website performance and increasing origin traffic on their extensive collection of sports apparel and fan gear retail sites. Akamai helped them improve their website performance and scale their platform by retaining search and social analytics, which provide essential insight on advertising performance and buyer behavior. Now, the last example was a major consumer electronics company, and they used us in order to prepare for massive demand around a mega sales event they had last year. They implemented a third-party waiting room solution to provide a great experience for their users to protect their origin infrastructure from a surge of traffic. They ran into problem with that third-party waiting room solution because it was run on the client, right?
The client was being circumvented by bad actors, which allowed the users to basically skip in line. They skipped the queue. They drove an unsustainable surge in origin traffic. Ultimately, that customer had to take down the sales event. What they did was they moved that third-party waiting room onto Akamai using our EdgeWorkers solution. Because it was running on the server instead of the client, the bad actors were no longer able to circumvent the waiting room logic, allowing the customer to control the traffic that was allowed into the checkout flow, and ultimately ensuring the sales went off without a hitch and the users got a great experience. All right. That's just the tip of the iceberg in terms of growth and compute on Akamai.
I told you all the way in the beginning that I'd cover both edge compute and 5G a little later, and here we are. The first area of growth in edge compute is around the objective of empowering our customers to move more and more complex workloads out to the edge. We intend to empower our customers with data services and expanded compute capabilities at our edge, enabling our customers to build and offload portions of their application back-end right into the cloud, really focusing on the use cases that can benefit from low latency, high reliability, and compliance. The second part Tom spoke about earlier, which is building a rich app ecosystem that enables our customers to solve any use case at the edge with low friction.
Within the app ecosystem, we'll build partnerships with third-party developers so that their products are going to be integrated at the Akamai edge and available seamlessly to our customers. We're really focused here on building an enterprise-grade app ecosystem where applications have been vetted and then curated to provide value to our enterprise class of customer. I also promised to talk a little more about how 5G is going to drive growth for Akamai. First, I thought it's important to understand what 5G will do in terms of improving performance at the last mile. So latency is lower, right? Throughput's going to be higher. A lot more people are going to be connected. More devices like IoT devices will be connected. More compute is going to migrate right to these edge locations. But you only get those benefits if you're at the edge.
That lower latency is only available at the edge at the 5G node itself. It means you'll need an edge platform like Akamai even more than you do today to take advantage of these new 5G capabilities. Just like broadband, 5G will increase the utilization of everything from mobile and home media, making AR and VR and cloud gaming possible, driving hyper-local in-venue content, in-car entertainment, massive explosion of IoT devices. All of these will drive traffic consumption up and edge computing, and that means real edge computing that happens at the edge, not in the core, to migrate to that new true 5G edge where Akamai is already today, and we're going to be there when 5G becomes a reality. In fact, Gartner predicts that by 2025, three-quarters of enterprise-generated data will be created and processed at the edge outside centralized data centers or the clouds.
That was only 10% in 2018. This could be one of the reasons why Gartner had said the edge is going to eat the cloud. While the better the last mile gets, the more that you need a true Edge platform. There's one last area that I felt was important to cover prior to closing my section today, because I get a fair number of questions about this every time I do any session. It's emphasizing Akamai's commitment to the developer experience and our DevOps capabilities. Akamai's role in our customers' technical architectures continues to deepen and widen. The Akamai Edge is a part of our customers' technical design and a key component of the automation they build for their businesses and their operations. We're supporting and embracing that with a continued investment in our product and platform capabilities for developers and DevOps processes.
Akamai for DevOps has resulted in a collection of powerful product features, tools, improved documentation to make it easier and faster for all of our customers' developers to harness the power of the edge in an automated fashion. Some of the things to focus on are things like better performance and usability across the full set of APIs, with a specific focus on content delivery APIs for speed and responsiveness. A new unified way to interact with Akamai via the Akamai command line interface or Akamai CLI. Third-party integrations with the most common developer and operations tools like Terraform and Jenkins. Integrated and cohesive documentation and code examples that show customers how to work across the Akamai experience and also gain inspiration based on what other developers are doing with Akamai.
The way we've worked on this has already transformed how developers work with us today, but we're just getting started on the DevOps experience. All right. Before I leave you, I wanted to remind you just one last time why Akamai continues to win in the marketplace. First, it's our unparalleled global reach and scale on demand. We are trusted and cloud neutral. We have a broad integrated product portfolio. We offer advanced edge-based serverless capabilities. We have best-in-class reliability. We drive conversion rate improvement. We provide real user monitoring and industry-leading quality of experience. Of course, it's our people. We have a global team of amazing OTT application and compute experts ready and willing to help. When it comes to CDN security and edge computing, Akamai is really the only solution for enterprises worldwide.
With that, I appreciate your time, and I look forward to taking your questions later. I'll leave you for your 10-minute break.
Hi, everybody. It goes without saying that we are living in a time of unprecedented change, but even notwithstanding the pandemic, I think it's safe to say that information technology is currently in a phase of extraordinary rapid and impactful change. What I want to do is talk about those changes and try to explain how it is that these changes are giving rise to the power of the Akamai Intelligent Edge platform. Let's start by looking at a quote from Gartner. Here's what they say, "Digital transformation and the adoption of mobile, cloud, and edge deployment models fundamentally change network traffic patterns, rendering existing network and security models obsolete." It's this last phrase I want to emphasize, "Rendering existing network and security models obsolete." To understand that, I think it's worth starting by looking at those existing network and security models.
Let's look at the traditional deployment model. The traditional deployment model that we've seen now for decades involves deploying performance and security stacks in a data center where the applications are. Applications live in one or maybe a small handful of data centers, and those applications can be accelerated and protected, or could be protected through the deployment of appliances in those data centers. Those appliances could include firewalls, gateways of various kinds, intrusion protection, intrusion detection, a myriad of appliances that could be deployed in the data center to deal with performance and security. That model, I think, made some sense back when all of the applications lived in one or a handful of data centers, but now we're living in a time when applications are moving to the cloud. In fact, applications, of course, have been moving to the cloud for quite some time now.
Let's look at cloud migration. As applications move to the cloud, it's completely natural then for the vendors who used to offer their performance and security capabilities in the form of an appliance to now pivot and offer their solutions in a virtual form factor, in effect, a virtual appliance that can then be also deployed in the cloud. That makes some sense. I claim that the cloud is not the right answer. Now, the cloud is the right answer for a whole lot of things. The cloud is a very important part of the technology ecosystem, but it's not the right answer for this. Why not? Well, it's not where the users are. It's not where the threats are. It's not where the applications are. It's also, by the way, not where the bandwidth is.
We heard about that from both Tom and Adam. It's not where the bandwidth is. It's not where the capacity is. I'm not going to spend time talking about that aspect. I want to focus on users, threats, applications. Let's start with users. Users are everywhere. If we consider, for example, employees, well, employees work remotely, and they work from everywhere. Certainly now in the pandemic, we have employees largely working from home, of course. Even before the pandemic, there was a reality of remote work even then. On any given day, employees in any reasonable size enterprise would be working from all kinds of locations, on the road, in an airport, in a coffee shop, at a customer site, in a hotel room, maybe from home.
Now as we look forward and we think about the pandemic ending and people going back into the office, sure, that will happen. I think we've reached a point now where employees expect that they can work from anywhere, and they will work from anywhere. Users are everywhere. How about threats? Threats are also everywhere. Threat actors come in many different types, from organized crime to advanced persistent threat groups, insiders, hacktivists, script kiddies, many different types of threat actors. One thing that they all have in common is they generally work through bots. They use bots to actually implement their attacks, and these bots can live anywhere, and they do. Bots are everywhere, and so these attacks come from everywhere. How about applications? Well, applications also are everywhere, and that wasn't the case years ago.
A decade or more ago, applications really did, by and large, live just in one or a handful of locations. That's not the case anymore. Nowadays, you're going to see applications across multiple clouds. In fact, even an enterprise that is consolidating on one cloud is probably still using multiple clouds. Certainly, if you're doing any amount of mergers and acquisitions, there's multiple clouds. Of course, there's going to be multiple cloud locations. Any enterprise that is doing business on a global scale or on a fairly large geographic scale is going to be using multiple cloud locations. Enterprises also still have legacy data centers. For some things, legacy data centers still make sense, and they're still there. They still exist, and many enterprises have significant legacy data centers, and many of them have no plans to eliminate them anytime soon. Also, office buildings.
Applications live in office buildings and in unknown locations. It's a fact of life that there are many unknown applications that IT just doesn't even know about. These could be applications that an engineer has running under their desk in some office somewhere. The scary thing is that some of these applications could be mission-critical. That's just reality. The applications are everywhere. That begs the question, can you really replicate your performance and security stack in all of those locations? I'm obviously asking that question rhetorically. The answer is obviously no. Given that you can't have a performance and security stack in all these locations, that means you're stuck in a world where you have to backhaul, and backhauling traffic destroys performance. Because think about what you're doing.
You're taking the traffic flow from the user, sending it across one or more networks to the performance and or security stack. From there on to wherever the application lives, you have that tromboning effect, and this destroys performance. It's also a disaster from the point of view of reliability and cost. Not a good thing to be doing. Backhauling attack traffic, well, that's even worse. Think about it. Taking all of that attack traffic and literally backhauling it into your virtual or physical data center where your security stack lives. That is a recipe for disaster. Any reasonable-sized attack or reasonably sophisticated attack is easily going to be able to overwhelm those defenses, maybe even overwhelm the network before you even get to those defenses. That has the effect of potentially bringing your entire enterprise down during the period of the attack.
That obviously is a true disaster. When you put all this together, there's only one reasonable solution, and that is the edge. Instead of backhauling traffic, instead move the performance and security stack to the edge, where it's ubiquitous, where it's near the users, near the threats, near the applications. That's exactly what we do with the Akamai Edge platform. Let's recall what is and where is the edge. This is a slide that Tom showed you earlier. The edge is the part of the Internet where you connect to all of those digital touchpoints, the devices, the users, everything. It's basically right at the last mile. It's where the capacity is, it's where the users are, it's where the threats are, it's where the applications are, and so on.
If you look at Akamai's programmable and Intelligent Edge in numbers, you see the following, and you've seen these numbers elsewhere. 4,100 + locations, 1,400 + networks, almost 1,000 cities and 135 countries. I want to focus on the city counts here, almost 1,000 cities. Ask the question, why are we deployed in so many cities? Well, the answer is because if you deliver traffic from one city to another, that means that traffic is traversing one or multiple network backbones. Network backbones are an expensive and relatively constrained resource. There's orders of magnitude, less capacity and more cost in those backbones than what you find at the last mile in the metro areas.
In fact, I think it's safe to say that the reason the internet performs as well as it does, or certainly a major reason why the internet performs as well as it does and has scaled as well as it has, is because Akamai takes so much traffic off of those backbones. Let's go to the next slide and look at what happens in a city. In a city, we oftentimes are deployed in multiple locations, and this is unique. If you consider what some other CDN or provider would do, if they're even in that city at all, they're going to be in one data center. We're going to be in multiple. That begs the question, well, why are we in multiple locations in a city? The first answer is access to providers.
We want to be inside all of the provider networks that connect to the devices, to the eyeballs, to the things, everything that is in that metro area. That generally means multiple providers. The reality of the way they provision their capabilities is that you can't reach them all. You can't be in all of those providers from one data center. You need to be in multiple. The next reason is a reliability reason. We want redundancy. We don't want to be dependent on one data center. If all of your assets, if all of your servers are in one data center, well, there's a real vulnerability there. You could spend a lot of money to try and make that data center very reliable.
You could purchase redundant power supplies and all kinds of capabilities to try and make that data center as reliable as possible, but it's never going to get all the way there. There's always going to be the possibility, and it happens, that that data center could go down. You need to be in multiple data centers for reliability. By the way, the combination of these two things, the fact that we're in multiple providers and that we have all this redundancy, is why you oftentimes see examples where a provider or a data center has a problem, and you see that other providers are much more impacted than us. In many cases, we're not impacted at all. We're almost always way less impacted than anybody else.
Now, there is a potential criticism of this approach, which is that some might say that this approach is inefficient because we're potentially storing content in lots and lots of different places and spending more money. That's not the case with us. With us, we're using metro area networks, and we have built intelligent software to effectively stitch these physical data centers together to act as one virtual data center. They act as one virtual data center. We get the efficiency of one data center with all of the redundancy and access to providers of multiple physical data centers. Now I want to conclude my remarks about the edge by talking about the rise of the edge.
The point I'm making here is that the macro trends that are out there in the world, out there driving information technology, those trends are only serving to make the edge more important as we go forward into the future. These trends that I've listed here are all durable trends. These are not trends that are here today, gone tomorrow. These are durable trends. We've talked about 5G. In fact, both Adam and Tom both talked about 5G, meaning more devices, more connectivity, more bandwidth at the edge, and therefore a greater disparity between the edge and what happens in centralized locations. Again, more need for the edge. IoT, of course, just means more and more devices. Big data and AI means vastly more data traversing these networks. Digital transformation, remote work I talked about. Cybercrime. We all know cybercrime is not going away.
It's only going to get worse. All of these are trends making the edge more important. I want to shift gears and talk about the platform, the power of the platform. As I've been talking about the edge, I've really been focusing on the where it is that we deploy our software, where it is that we deploy our capabilities. When I talk about the platform, I'm focusing now on the what it is that our software does. The key point is that our software forms one edge platform to deliver performance, scale, and security across all data flows. To really unpack that and understand what that is, it's worth taking a look at the traditional model, which is siloed. The traditional model is siloed.
What I'm illustrating in this picture here, on the left, you see various categories of applications that an enterprise might have, from your public-facing applications to your private cloud applications, internet web applications, and so on. In the middle, you see some of the security capabilities that you would want to deploy to protect those applications. For example, with your public-facing applications, you probably want DDoS protection. You probably want a web app firewall, bot management. There's actually probably a whole bunch of other things you want. For simplicity of illustration, I'm only showing these three. Next on the list, you have our private cloud applications. For those, you probably want firewall. You probably want a zero-trust access model and many, many other things. Again, for simplicity, I'm just showing these two.
For your access to web applications, you probably want a secure web gateway and on down the list. What I'm illustrating here is that each of these security capabilities is generally delivered in a silo. It's generally delivered from multiple vendors. Even if it is multiple things from one vendor, it's probably separate appliances or virtual appliances. Again, the traditional model is siloed, these capabilities being delivered in separate appliances or virtual appliances. The siloed model is a problem. Silos lead to vulnerabilities. This is important to understand. Silos lead to vulnerabilities. Silos create complexity, and complexity leads to vulnerabilities. This, again, very important. Truly, complexity is the enemy of security. Silos also create multiple policies that have to be managed. That means that you end up with gaps. Silos leave gaps, and gaps is where the vulnerabilities live.
Here's the thing, you can't daisy chain edges. Well, okay, maybe in theory you could do it. You could try and do it. It's certainly not advised. It would lead to all kinds of problems with reliability and performance and things like that. Not a good thing to be doing. Now, in the old deployment model, in the data center or in the virtual data center, yeah, you could provision appliances, multiple appliances. Provision virtual appliances, multiple virtual appliances, daisy chain them together. Yeah, you still have the problems that I illustrate here, but you could at least do it. As we go forward, you just can't even do it, and it's just a really bad model. Silos is a really bad model going forward. It's not just me or Akamai saying this is what customers are saying, and this is what analysts are saying.
Enterprises crave simplicity and unification. Let's look at this quote from Gartner. Gartner says a remote or mobile user is going to need access to the internet. They're going to need access to various SaaS applications, for example, Microsoft Office 365, Salesforce, Box, maybe numerous others. They're going to need access to enterprise private applications all at the same time. Gartner rhetorically asks the question, are these really different problems? Obviously, the answer is no, and they're making the point that going forward, it makes no sense to deliver these capabilities in silos. These capabilities should be delivered as one platform. Let's take a look at another Gartner quote, which gets to the same point.
By 2023, more than 30% of public-facing web applications and APIs will be protected by cloud web application and API protection services, which combine distributed denial of service protection, bot mitigation, API protection, and WAFs. Combining. They're making this point that it makes no sense to deliver these capabilities in silos. These capabilities should be delivered as part of one platform. When you put all that together, again, the solution is pretty manifest. The solution is a single edge platform, the Akamai Edge Platform. One edge platform for performance, scale, and security across all of your data flows. With this comes tremendous power. There's tremendous power in one platform. The one platform gives you simplicity. It unifies policy across all of those flows. It eliminates the gaps. It gives you something that I like to call the power of multiplicative security.
Let me explain what I mean by multiplicative security. It's the idea that as we develop new capabilities or improve existing capabilities, all of the data flows on our platform can benefit from that. For example, if we improve our web app firewall, or we improve our bot management or anything else, all traffic flows that go through the Akamai platform are able to benefit from that. Contrast that with what would happen in silos. In silos, if one silo gets better, it doesn't do anything for the other silos. The power of multiplicative security, very powerful. I've been focusing here on security, but the same story holds true for performance. One platform for security and performance. The platform that secures traffic flows also accelerates them. In fact, I think Tom pointed this out earlier, it's really the same servers.
As your traffic flows through one or more Akamai servers, the same servers that are securing traffic also accelerates traffic. This creates a platform that means that our customers can easily consume, fill in the blank, as a service. Whatever it is they need, they can consume that easily off of one platform as a service. Whatever app and network security capabilities they need, from app and API protection to fraud protection, access control, infrastructure protection. These are capabilities that Rick talked about earlier. Whatever capabilities are needed, easy to consume, one platform. Edge applications. Adam talked about the edge applications. From EdgeWorkers to EdgeKV and so on. Again, easy to consume, one platform. Edge delivery. Adam talked about the edge delivery capabilities from Ion to streaming, software and gaming downloads, and the like. Again, all easy to consume as a service off of one platform.
This means that the platform delivers real business objectives easily. Consider the recent example now of local and state governments who are rapidly rolling out new web applications, new websites to control and manage the rollout of the COVID vaccine. They need a lot of different capabilities. They need visitor prioritization. It's unfortunate, but a truth that, yeah, they need DDoS protection. They need web application and API protection. They need delivery for scalability. They need acceleration for that experience, and more. They need a whole bunch of capabilities. They can get all of that on one platform with Akamai. Think of what it would be like if they had to do this through multiple appliances or virtual appliances through silos. You could stitch all those things together, but it would be a nightmare. Instead, one platform to deliver all of those business objectives.
The final comment now that I want to make about the platform is, again, the rise, the rise of the platform. The macro trends, as they are with the edge, are only serving to make the platform more important going forward. When you look at some of these trends, a lot of them are actually the same trends that are giving rise to the edge. You see a couple of different ones here. For example, vendor consolidation. We hear this from customers all the time. Customers want to consolidate on a smaller number of vendors. It makes sense. Look at the one on the right, though. The developer as the customer. This is actually a very important trend, a durable trend, that Adam talked about.
Think about the power of a platform when it's programmable in the hands of a developer, as opposed to silos, which may or may not be programmable. One programmable platform in the hands of developers is an extraordinarily powerful capability. Putting it all together, we clearly see the value and importance of an Edge platform. When you look at the Akamai Edge platform, we are operating at an unprecedented level of scale, performance, reliability, data, capabilities, the breadth of capabilities, the depth of capabilities. You can see the numbers here. I'm not going to go through all the numbers, because I think most of these numbers have been presented earlier. I think the point is clear, the Akamai Edge platform operates at unprecedented scale and capability. Of course, we're also continuously innovating on the Edge platform. We have a great culture of innovation, always have.
One of the ways I like to characterize our engineering culture is one of continuous learning. We're all continuous learners. We love learning new things. We love exploring the new technologies that's out there. When we find new ideas and new technologies, we use them. We bring them into the platform. In other cases, of course, we have to invent new things. We've done that again and again throughout our history. It's a great culture of continuous innovation on the Edge Platform. The result is that the Akamai platform is unprecedented, and it's an enormous advantage for us. The Akamai Edge platform is a huge advantage. Let's go to the next slide, which is my last slide. The Akamai Edge platform is extensible and modular. It's reliable and scalable, and it's programmable, and it's incredibly difficult to replicate.
I can speak from experience, by the way, because I've been here. I can tell you it was very difficult to build in the first place. We've done it, and we'll continue innovating. There's no question it would be a very difficult thing to replicate. With that, I thank you all, and I can now hand it off to PJ.
Great. Thanks, Bobby, and nice to meet all of you. I'm Paul Joseph. I joined Akamai shortly after the company went public, and I've been running our media and carrier sales organization for the last few years. I want to first recap the industries we serve and the types of customers we work with in each of these verticals. Today, we serve more than 850 retailers, 41 of the 50 largest hotels in the world, and numerous leading travel companies. These are enterprises like NIKE, LVMH, Alibaba, Marriott, and Delta Air Lines. We work closely with eight of the 10 largest financial services companies in the world, including companies like Mastercard, Fidelity, HSBC, and JP Morgan. We also count over 950 software and high-tech companies as customers, including Salesforce, Adobe, Nvidia, Oracle, and IBM.
We serve over 240 OTT and direct-to-consumer companies worldwide, including Disney, NBC Peacock, Sky, Discovery, Stan, BBC, ITV, and HBO Max. In gaming, we serve 23 of the 25 most popular video game publishers and all of the major console manufacturers. This list includes companies like Sony, Microsoft Xbox, EA, Riot Games, and Epic Games. Our public sector team works with over 200 government agencies, including all five branches of the U.S. military. We have over 275 global carrier relationships, a few of which we'll talk about later in the presentation. We serve nine of the 10 global healthcare companies, including Merck, Pfizer, and Eli Lilly, and we work with more than 50 automotive companies worldwide. Okay, this slide illustrates a few of the other global brands that leverage our services.
All of these companies invest in Akamai because of our scale and reach, our performance, and our consistent track record of reliability. It's also worth pointing out that many of these customers use our security services, and their industry peers benefit from the network effect of the Akamai Edge, which is an often-overlooked benefit of our platform. We see far more attack traffic than other CDNs and are able to build all of this intelligence into our security services in real time, which means we stop the latest threats continuously and earlier than our competitors. For example, we may see a very targeted DDoS campaign in a specific industry vertical, and when we do, we proactively integrate these learnings into our security offerings for the benefit of all other customers before these same attacks get to them.
Okay, now I'll talk briefly about how we segment and service our install base and our go-to-market approach for new business. We've done an extensive analysis of our customer base over the last 10 years to help shape our go-to-market strategy. We look at spend and product consumption patterns across all existing customers for all products. We review data about how these companies use our capabilities, why they're using them, their renewal history, how many services they have under contract, and their annual spend by year. We use this data to build and refine our coverage models to ensure we're aligning our cost of sale to customer value. I'll start with the SMB segment. Our SMB segment are generally clients sold through our carrier OEM relationships and our aggregator partnerships. We define the SMB segment as local or regional businesses with smaller IT and security budgets.
These are normally point solution buyers. They're low or no touch. This segment is channel-driven and where wallet share potential is less than $75K annually. We have a very low cost of sale in this segment and an extremely efficient customer acquisition engine for onboarding hundreds of thousands of customers. Our carrier OEM partners provide us with a tremendous amount of reach and offload in this segment, since they take on all of the marketing, all of the account management, and customer service work. Our enterprise accounts are a mix of national and global companies with 1,000 - 10,000 employees and growing security and digital budgets. These enterprises have wallet share potential of $75K-$250K annually. These are also lower touch customers that we serve direct with our inside sales teams or through local carrier and SI partners.
We don't see a huge amount of traffic variability with these customers, though we have a strong pipeline of targeted new product attach opportunities in this segment, particularly around enterprise security and our edge application offerings. Our large enterprise segment are customers and prospects with 10,000+ employees and substantial IT and security budgets. These enterprises have the potential to spend more than $500K a year with us. This segment is direct and indirect led, with strong engagement from our carrier and SI partners. This class of customers, coupled with our strategic accounts and our high-potential prospects, is where we have the greatest wallet share potential. Characteristics of customers in this segment include a history of successful renewals, strong traffic growth, and strong attach rates for our security capabilities.
A large percentage of these customers also leverage our professional services organization to ensure their web estates and subscription video offerings are fully optimized and secure. We are a preferred supplier in this segment but also have plenty of room to grow. Think of a big bank that is using Akamai to accelerate and secure some but not all of their consumer-facing or B2B sites, or a finserve customer that is leveraging our Prolexic DDoS services for a subset of their international data center footprint, but have plans to standardize on our platform globally. In situations like this, we have numerous opportunities for expansion. It could be an OTT customer that is comfortable outsourcing delivery to one or two CDNs and is in the process of launching streaming services in a geography where we have a significant performance and scale advantage over our competition.
Here we have great consolidation opportunities for traffic and an increasing amount of security upsell opportunities to prevent against things like credential stuffing, account takeover, and content piracy. A subset of our strategic accounts are also some of our largest revenue-generating customers with high utilization rates and several services under contract. Most of our high-value strategic accounts also have dedicated service and support teams. These are very sticky relationships for us. You can kind of map this segment to our internet platform customers, large retailers and social media companies, large SaaS and PaaS providers, government agencies where we're doing classified custom work, and finserve customers that are heavily invested in our security stack. Okay. This slide is designed to give you a view into our revenue diversity by customer spend and the balanced revenue contribution we have across the business.
In 2020, we had over 500 customers drive $1 million-$5 million in revenue for us. More than 50 customers generated between $5 million and $10 million in annualized revenue. Over 20 customers did between $10 million and $30 million of revenue, and we had eight customers do more than $30 million each with us in 2020. These eight customers generated roughly 15% of total revenue. Okay, I'll now talk briefly about our security penetration in a few key categories to highlight the opportunities we have to expand our security market share. Rick touched on some of this earlier in his presentation. We work really closely with our sales operations group to make sure that our sellers have a fresh view of the security white space in our base.
We also do the same type of analysis to identify new opportunities for edge application offerings, where we're currently seeing strong interest from enterprises looking to move their microservices to the edge. This slide illustrates the high-level output of that work for a few of our security services and highlights the white space in our top 500 customers. As you can see here, we have strong penetration of our market-leading WAF capabilities, we have plenty of room to grow our security footprint with our application and network security offerings. Less than 50% of our top 500 customers leverage Bot Manager, Page Integrity Manager, or our Prolexic DDoS services. The same is true for our secure web gateway and application access services. As credential stuffing, account takeover, and ransom DDoS attacks continue to increase, we stand to benefit considerably if we can execute on our cross-sell and upsell campaigns.
On the carrier side of things, roughly half of the top 25 carriers in the world buy our DNS infrastructure services, but even fewer are leveraging our secure business and secure mobile services from Asavie. There is a big opportunity to accelerate revenue in this segment, especially given the rollout of 5G and the work-from-anywhere trend that is here to stay. Our carriers want easily consumable security services to help safeguard their SMB and enterprise customers, and they want a simple way to create and automate private network services. Akamai enables them to do this easily while providing a seamless way to integrate these capabilities into their own service bundles.
If you look at what two of the world's largest carriers are doing with us today, Comcast with their Comcast Security Edge offering, and AT&T with their AccessMyLAN service, you can see why we're bullish on the network security segment as a driver of future growth. The last point I want to make on this slide is that we ended 2020 with over 200 customers doing more than $1 million in security with us, up from just over 100 million customers in 2018. Okay, I wanted to share a little more data on our revenue diversity. In terms of revenue by geography, 56% of our revenue in 2020 came from North America, and 44% came from markets outside of the U.S., up from 31% in 2016. We plan to continue to leverage our scale and performance advantage internationally to accelerate revenue in non-U.S. markets.
It's also worth pointing out that we have very advantageous network economics in international markets, given the volume of traffic we serve outside of the U.S. It's very hard for our competitors to match our capacity and the platform leverage we have given their limited footprint outside of the U.S. We're currently seeing excellent traction and growth in emerging markets like Southeast Asia and Eastern Europe, and in LATAM, where our Exceda acquisition has provided us with a very solid foundation to grow our market share in the region. Because of these trends, we expect our international mix of revenue to increase considerably over the next three-five years. I now want to say a few words about our newly unified sales organization.
We don't expect any customer impact or disruption from the reorg, as account-level assignments will largely stay the same, and customers will see the same faces from a sales and support standpoint. We do expect this new structure to make us more efficient in terms of how we respond to the needs of our customers and partners. In terms of the scale of the selling organization, today, we have feet on the ground in 85 cities and 31 countries, very close to our customers. Our field sellers are supported by a tenured pre-sales organization with deep vertical industry expertise. Both organizations are highly incented to sell our whole portfolio, including all of our security products. We also have a specialist group inside our pre-sales team that is hyper-focused on EdgeWorkers and edge application adoption.
The field is also supported by an advanced technology group of security specialists that focus on EAA, ETP, and Prolexic. These are experts on enterprise security and zero trust networking who augment our pre-sales teams and sales organization to progress opportunities faster and to accelerate our time to revenue. We also have dedicated partner account executives mapped to all of our partners in each of the major geos. We work very closely with our marketing organization, where we have a very capable team of SDRs that nurture and qualify inbound and outbound leads for our field sellers. You'll hear more about the marketing strategy in a few minutes from Kim. Now turning to our partner program, we have four major categories of partners. We have strong go-to-market partnerships with many of the largest carriers in each of the major geos.
Many of these carriers are full portfolio resellers, while others OEM our access control capabilities to serve their SMB and enterprise customers. We also have strong SI partners and a robust ecosystem of VARs that bundle our capabilities with their content management offerings. We have several productive aggregator partners who embed our capabilities into their infrastructure and PaaS offerings. I now want to make a few more comments about our carrier relationships. This is a big differentiator for Akamai, and none of our competitors have the kind of deep, multifaceted relationships that we do with carriers. Today, we have 1,400 network partners and strongly aligned interests around cost management and performance. We also count 300 carriers as customers. Many of these carriers use our acceleration and security services for their own B2B and consumer-facing websites.
Several buy our managed and licensed CDN offerings to power their owned and licensed content offerings, many use our DNS infrastructure services to power their broadband offerings. On the reseller side of things, we have carriers like AT&T, Telstra, Singtel, BT, NTT that all represent our whole portfolio to their enterprise customer base. We inherited some excellent 5G partners as a result of our SBE acquisition, all of who are seeing strong enterprise demand for private network services. We'll be working hard over the coming months and years to expand this partner ecosystem. Lastly, I wanted to quickly cover some additional partner program data points. Akamai generated roughly $1 billion in total revenue from partners in 2020, over a third of this came from our carrier partners.
41% of total company bookings came from partners last year, and 69% of all new customer bookings came from the channel. We ended 2020 with 3,000 discrete indirect customers. We also onboarded 300,000 SMBs onto our DNS security platform through our carrier OEM relationships. We are seeing very strong traction with our new partner program that rewards sellers in multiple ways for investing with Akamai. In closing, we're very excited about the numerous opportunities we have in front of us to expand our business. I'll now turn things over to Kim for a high-level overview of how the marketing organization is helping sales drive demand. Kim?
Thanks, PJ, and good afternoon, everyone. I'm Kim Salem-Jackson, Akamai's new CMO, and I am thrilled to speak with you today about the work we're doing within the global marketing organization to support both our go-to-market and thought leadership efforts. Let's dive in. The goal of marketing at Akamai is to be a growth engine for the business and maximize our impact. Our vision and mission are simple. We want to solidify Akamai as the leader in cloud security solutions in edge applications, while accelerating revenue growth by both attracting and engaging key buying teams and influencers. Akamai has a large and growing market opportunity. Our portfolio and customer base are highly diversified, providing us opportunity to both acquire new customers and cross-sell adjacent solutions to our base.
Marketing will draft off of our leadership position as a trusted partner to the world's leading enterprises as a category leader in DDoS mitigation, web application firewall, bot mitigation, and Zero Trust. As the world's largest and most distributed edge platform to address what we see as a $47 billion market opportunity. How we capitalize on the market opportunity will be achieved by harnessing the power of our trusted brand and data-driven ecosystem. From a brand standpoint, Akamai has a long history of reliability and trust among our customers, which will act as a catalyst for accelerating our edge security and edge application growth. Our data-driven ecosystem is fueled by using AI and machine learning to anticipate customer needs and meet customers where they are in the buying journey, essentially shrinking the timeframe from interest to revenue.
The analogy I like to use to describe this motion is we now have the ability to market to the square inch versus the square mile, maximizing our marketing dollars to their fullest extent. You won't see us show up everywhere, you will see us show up where it matters most. Our trusted brand, coupled with our data-driven ecosystem, work in concert, enabling us to deliver a superior customer experience. Let's talk about our brand. Akamai's trusted brand equity is demonstrated in multiple ways, from awareness to purchase. Our category leadership is showcased in our 200% year-on-year growth in Google searches for Akamai Security. This data point illustrates pure inorganic growth. People are talking about Akamai Security and aggressively consuming our content. Akamai has won the Gartner Peer Insights Customers' Choice in 2021. It's based on customer reviews, receiving a 4.6 out of 5.
Trust in Akamai is further validated by our customers. According to research, we possess strong brand equity, with eight out of 10 customers rating us as excellent and likely to consider for purchase for our security solutions. We are committed to growing that same level of awareness, specifically in zero trust. It's backed by the confidence of analyst firms such as Forrester. You have heard from my colleagues about the breadth of global brands we are honored to call our customers. What is critical to point out here is that many of these organizations are also the most innovative companies, such as Disney+, Airbnb, and Microsoft, to name just a few. Let's not forget, many of these organizations acted as a critical lifeline to us all in the last year during the pandemic.
Now let's talk about the key pillars to our growth engine, starting with new logo acquisition. New customers have been and will remain a high priority for marketing. We utilize predictive data to identify the right prospects for Akamai, yielding a strong lift in our conversion rates. We are also committed to penetrating our white space and driving security cross-sell, particularly for the zero trust products. We are actually seeing the average deal size grow 2x for accounts highly engaged in marketing. As you heard from PJ, international growth is a priority, and we have identified our highest growth markets and are committing to significantly increasing spend there to accelerate our growth trajectory. Last but not least, marketing continues to lean in on the channel as another key growth driver. We see a 3x probability a deal will close when we work with channel partners.
We are so committed to our growth pillars and strategy, we actually compensate our field marketing organization on sourced revenue, not pipeline, to keep us focused on top-line growth. As a reminder, all these pillars are powered by our data-driven ecosystem. Moments ago, you heard PJ provide a detailed overview of our sales motion by business size. Marketing will draft off this approach and align our best marketing strategies to further penetrate the market segments. At the top of the pyramid, we will focus on high touch, highly customized account-based marketing program. As we work our way down the pyramid, we will deploy marketing tactics utilizing predictive data focused on penetrating the large and mid-enterprise segments. Lastly, we will target the SMB with a channel-first strategy and channel-first marketing programs.
We have spoken a lot about the data ecosystem. I want to showcase some of the intelligence we are now able to both derive and action from the data. From marketing source deals, we understand precisely how many times we need to touch an account before we close the deal for both new and existing customers. For example, we know it takes approximately 40 marketing touches to win a deal. When we see accounts not engaging enough with us, marketing leans in. We know by deploying predictive marketing strategies, we achieve a 5x lift in our performance. As a result, in a year where most organizations saw a decline in their new customers, we grew more than 20%. Last but not least, using our customer lifetime value calculation, we have achieved a 46% increase in ROI by deploying these data-driven strategies.
Why do you care about all these stats? Because in collaboration with PJ and his global sales organization, we want to convey the message that you can count on our combined go-to-market efforts to make smart investments and maximize our business impact. In closing, we have demonstrated our brand leadership through a long history of trust validated by our customers. We have accelerated our growth trajectory for market categories such as Zero Trust by harnessing our data-driven ecosystem. These two pillars working in concert ensure we can deliver a superior customer experience. I want to thank you so very much for the pleasure of your time. Now I would like to introduce a short video on our work in ESG featuring both Nicole Fitzpatrick and Khalil Smith.
My name is Nicole Fitzpatrick, and I am Akamai's Deputy General Counsel, and in conjunction with Akamai's newly formed Environmental, Social, and Governance office, our Chief ESG Officer. More on that in a moment. First, I want to talk to you about Akamai's values. Customer first, one Akamai, innovation, inclusion, urgency and persistence, integrity and trust, and giving back. These seven values have continued to form the bedrock upon which we've made our decisions for years. They have informed not only what we do, but how we do what we do and why we do what we do. These values have enabled us to weather market forces and competitors while maintaining a focus on our core business, prioritizing our customers, advancing the right M&A opportunities, and delivering strong stakeholder returns. Not every conversation at Akamai begins and ends with these values. Every decision takes them into account.
Those same seven values have influenced our work around sustainability, our Akamai Foundation, and our Office of Inclusion, Diversity, and Engagement. Along the way, each of those teams have done incredible work and have been recognized by industry experts as leaders. Today, we are introducing our new ESG office, and in doing so, we are making it clear that we are putting our efforts behind moving from being leaders to being true catalysts for systemic and positive change. By establishing a formal ESG office and thereby formulating cohesive strategies for measuring, evolving, and communicating our progress, our goal is to drive growth, embed sustainable business practices, and cultivate the world's best culture. We also know that through sharing our wins and lessons transparently and without reservation, we will inspire other companies to do the same.
You, as investors, will continue to have a front-row seat through that enhanced disclosure into how we are using our dollars wisely to invest in programs that create a virtuous cycle of business health and resilience. To provide a brief example of what I mean, I've invited Khalil Smith, our VP of Inclusion, Diversity, and Engagement, to talk about Akamai's Technical Academy. Khalil?
Thanks so much, Nicole. It's great to be with you all. Akamai Technical Academy was launched in 2016 as a training program designed for people who have an interest and aptitude for technology but may not come from a traditional technical background. The program consists of five paid months of classroom training, after which participants are placed in a variety of roles across the organization with the intention to convert them to permanent employees after a minimum of six months. We have hired scores of incredible talent. To Nicole's point about inspiring others and catalyzing change, the program has been so revolutionary and so successful that Harvard Business School now teaches it as a case. This is an example of a program that lives our values, benefits our business, invests in our communities, and diversifies our talent.
Tom, Nicole and I could talk about our program and our people for hours, but we'd love to turn it back over to you.
Thank you, Khalil, and of course, Nicole, for that very insightful look on ESG. We'd now like to provide you another 10-minute break, and when we return, we'll have our Chief Financial Officer, Ed McGowan, provide you a financial overview. See you in a few minutes.
Good afternoon, everyone. I will be the last presenter of the day, and then we'll jump into Q&A. First, a few quick housekeeping items before I start. First, we've introduced a lot of data and some new metrics today. Most of what you have seen will be updated on an annual basis and not something we will share every quarter. All of the materials we share today will be made available on our investor relations section of our website. Additionally, I'll discuss later in my presentation about our use of capital and how M&A has always been a part of our core strategy around innovation and growth.
The numbers you have heard or about to hear assume some level of M&A activity. I would characterize this assumption to include a level of M&A activity that is roughly consistent with our track record over the past three - five years. More significant acquisitions could potentially result in a shorter-term impact to margins and/or a longer-term impact to revenue growth. We would plan to call that out at the appropriate time if we were to do such an acquisition. Of course, we plan to maintain the same level of discipline as buyers that you have come to expect from us. Finally, in a few slides, I will be sharing relative profitability metrics for our Security Technology Group and our Edge Technology Group. As Bobby noted, all of our products run on a shared platform, and the two businesses cannot be separated.
As a result, in order to arrive at relative profitability for each group, we had a significant amount of cost that we needed to allocate, including G&A, sales and marketing, cost of goods sold, cost to run the platform, and CapEx. We leverage a number of models to allocate these costs, and we believe the numbers we will share in a few minutes are a reasonable approximation of the relative profitability for each business. First, a quick recap of what you've heard today. You heard Tom talk about our leadership position across cloud security, edge delivery, and edge computing, as well as introducing a three to five-year goal of 20%+ security revenue growth. Rick provided a deeper look into our diversified security portfolio and highlighted several exciting growth areas, including application and API protection, fraud protection, and access control, just to name a few.
Adam laid out his strategy for our industry-leading capabilities in edge delivery and highlighted a very fast-growing edge applications product suite. Bobby walked us through the power and differentiation of our unique Intelligent Edge platform. PJ and Kim talked about our go-to-market strategy and the focus on cross-selling and security-led new customer acquisition approach. Nicole and Khalil talked about some of the amazing work we're doing around ESG, that highlights one of the key pillars of our company's culture and values. Today, I'll try to pull all of what you heard together this morning into what it means from a numbers perspective. I'll start with some key trends from the past few years, focusing on revenue and profit growth, revenue diversification, and margin expansion.
Next, I'll provide a deeper dive into the Security Technology Group and Edge Technology Group. Again, looking at revenue growth drivers and relative profitability of the two groups.
I'll then talk about our three-five year revenue and margin goals, and wrap it up by talking about capital deployment in M&A. Over the past five years, we have added nearly $1 billion of revenue, resulting in an 8% compounded annual growth rate for the company over that period of time. We more than doubled our security revenue over the past five years to over $1 billion in 2020, a 33% compounded annual growth rate. Truly incredible. As Tom mentioned on our Q4 earnings call, we are one of very few companies that generate more than $1 billion of annual revenue growth from cybersecurity solutions. Finally, while Edge Technology Group has been growing slower than security, Adam highlighted several exciting growth opportunities, including edge computing, 5G, IoT, and continued traffic growth from OTT and gaming.
On the last slide, I talked about how we added nearly $1 billion of revenue over the past five years, and while revenue grew at 8% compounded annually over that period, we grew non-GAAP EPS twice as fast at 16% compounded annually, and we delivered over $5 a share of non-GAAP EPS in 2020. Over the past several years, we expanded our non-GAAP operating margin from a low of 24% in 2017 to 31% in 2020. I'll talk about how we accomplished this margin expansion in more detail in a few minutes, but it's important to note that we did this while still investing in R&D to enable future growth. One key element that distinguishes Akamai from many of our competitors is our revenue diversification. You can see from this chart that our revenue is well diversified across product, geography, and customer.
For example, today, security is over 1/3 of our total company's revenue. PJ talked about our investment in go-to-market resources outside the U.S., and Adam and Tom talked about how having as much traffic as we do enables us to get cost-effective edge network deployments around the world. These factors, great people, and a vast network outside the U.S. gives us a huge competitive advantage internationally, and you can see that resulting in a very rapid international growth over the last five years. From a customer perspective, PJ mentioned we have over 500 customers that pay us more than $1 million annually. We have only eight customers that generate more than 1% of total revenue. I often get asked about the potential of our large OTT customers moving their CDN needs in-house.
As you can see, if that were to happen, any one customer shifting business in-house would not have a material impact on the business. As we've demonstrated over the years with our internet platform customers, all of whom have their own do it yourself CDNs, there's still a place for Akamai in a do it yourself world. Before I move on to a deeper dive on the security and edge groups, I wanted to spend a minute on our margin expansion efforts over the past several years. It's important to point out while we're scaling our operating expenses, we're able to continue to maintain a very healthy growth in research and development over that same time period. You can see from the slide we expanded our non-GAAP operating margins by 7% from 2017 to 2020.
This was accomplished by a combination of back office and IT efficiencies, the establishment of a formal procurement function that enabled us to save millions of dollars in third-party spend, continued innovation from our platform organization to improve server efficiency and drive down the cost to serve traffic, increased channel leverage in our go-to-market function, along with much more efficient marketing spend that Kim talked about earlier. You can see, gross R&D spend over the timeframe remained at a healthy 14% of revenue during that time. A reminder, approximately 50% of our gross R&D spend is capitalized in accordance with GAAP, and then it's amortized through our depreciation line. The best way to look at gross R&D spend is to look at R&D expense plus the capitalized software line we record in our CapEx. Let's take a deeper look at our Security Technology Group revenue.
As we look under the covers of our $1 billion security business, you heard Rick talk about a number of fast-growing product lines, including application and API protection, which includes our flagship Kona Site Defender or WAF product line. That group grew at 28% in 2020, and those products ended the year with $480 million of revenue. Rick also highlighted our goal is to continue to grow that business at a rate of 18%-22% over the next three - five years. Fraud protection, which includes bot management, ended the year at $176 million, up 36% year-over-year. In another area of very exciting new product development, we're expected to see very strong growth over the next three to five years.
Our access control products that include zero trust and carrier security products like Asavie, grew at 62% in 2020. Tom mentioned on our Q4 Earnings Call two weeks ago that we hope this product suite will approach $100 million in 2021. If we achieve this goal, it would mean roughly doubling revenue in this category from 2020. These metrics may look familiar to many of you as we typically provide these stats in our quarterly earnings Q&A section. Earlier, you heard PJ and Kim talk about our efforts in sales and marketing to cross-sell security into our install base, along with a security-first new customer acquisition strategy. You also heard PJ talk about the fact we now have over 200 customers who spend more than a million dollars annually with us on our security products. The team has done an excellent job cross-selling security into all of our key verticals.
The number of customers that have purchased at least one security product has increased from approximately 50% in Q1 2019 to 62% at the end of 2020. Over that same time period, the number of customers buying two or more security products has increased from approximately 25% in Q1 2019 to approximately 31% in Q4 2020. Despite the fact that over 60% of our customers buy a security product today, we believe there's still a long way to go in our install base. Both PJ and Rick talked about the security penetration in our top 500 customers and highlighted the great work that our sales and marketing teams are doing with this group of customers. This slide shows a look at our security penetration by some of our main security products across our entire customer base of approximately 6,500 customers.
You can see, for instance, that only about one-quarter of our customers today purchase our web application firewall products, our largest revenue-generating product. Page Integrity Manager that was just launched last year is starting to gain traction with our install base and becoming a larger percentage of our security bookings. You can also see that our EAA and ETP, or Zero Trust products, are continuing to gain momentum across the install base. Despite the fact that we've developed a number of tools that enable our customers to run our security products on their own, we're finding more and more customers are asking us for our security services experts to act as an extension of their internal security teams to help them manage the ever-evolving and complex threat landscape. Now let's move on to the relative profitability of the Security Technology Group.
Just as a reminder, all the products run on the shared platform, as I just mentioned, and the businesses really can't be separated. We do a significant amount of allocation of costs, including our G&A sales, network costs, and our CapEx. As you can see here, this business is very high gross margins, and we estimate the non-GAAP operating margins of this group to be in the mid-30s. This business is also much less capital-intensive from a network CapEx perspective than our CDN business. Some of the costs that are included in our cost of goods sold would be an allocation of things like bandwidth costs and co-location costs. We do have some direct costs, like third-party cloud, and then the costs of our security services team that's included in our cost of goods sold as well.
As for network CapEx, most of the spend is an allocation of the network. Remember, the bulk of the network assets are used to deliver traffic for our large media customers. However, we do have some direct security product CapEx, like our scrubbing center costs associated with Prolexic services. You can see here that our capitalized R&D is running at about 8%-10%. As I pointed out earlier, you should really think about this as part of our gross R&D spend. Finally, you can see that we are operating very close to our target non-GAAP operating margin levels today. We believe it's important to keep investing in R&D to capture the exciting growth opportunities we see in security. Now let's shift gears to the Edge Technology Group or CDN business. You heard Adam break this business down into three main areas.
The first being Edge Delivery, the largest, about $1.7 billion in 2020, which grew at low single digits. Our Edge Application business, which was approximately $150 million in 2020, and grew at approximately 23% last year. Our Services business, which was approximately $270 million in 2020. One way to look at this business that I think is very helpful is to look at it through the lens of our key verticals. You can see here that about 63% of our business in 2020 came from media, gaming, and high tech. These verticals represent over 95% of our traffic. These customers tend to have lower gross margin as we face volume-based pricing pressure from these customers, and the primary technology used to deliver this traffic is caching. The remaining verticals are more in line with what we used to see in our former web division.
Customers in verticals like financial services or retail, those customers tend to push less traffic, have higher gross margin, the primary technology used to deliver this traffic is application acceleration. It is worth noting that about 16% of our CDN revenue comes from retail, travel, and hospitality. We have talked about on our last few earnings calls that these two verticals are the hardest hit by the pandemic. As a result, we expect to see continued pressure from these customers, especially around renewal time. That we would expect it would negatively impact our growth rate in 2021, as we discussed on our last earnings call. As Adam noted, we probably don't see a turnaround until the back half of 2022 in these verticals.
Finally, included in other verticals are things such as manufacturing, business services, automotive, pharmaceutical, and education. One of the main growth drivers for the Edge delivery business is traffic. Adam talked about the record traffic we saw on our platform in 2020, driven by significant growth in both OTT video and gaming. Bobby, Tom, and Adam all talked about the challenges of delivering these kinds of volumes from the core internet data centers. It's just not possible to deliver HD-quality video streams from a handful of locations. That is why we believe our Edge network puts us in a unique position to capitalize on OTT and gaming tailwinds in the coming years. As you can see from the chart, OTT video, which is our fastest-growing segment of traffic, now represents over half of the traffic we see on our platform.
Here, you can see the three product revenue categories I mentioned a few minutes ago. We just talked about delivery. Now let's spend a minute on our rapidly growing Edge application business. You heard Adam refer to our platform as the world's most distributed serverless compute platform, that every day our customers leverage products like EdgeWorkers that enables our customers to deploy code without having to worry about scale, or our EdgeKV product that enables customers to leverage our Edge for local data processing. There's the potential for all the new use cases for our Edge applications as 5G becomes more pervasive and billions of IoT devices come online. 5G will enable increasingly more and more low latency sensitive workloads, and our Edge network is the perfect place for those workloads to run, and we can do so at very attractive economics.
Moving on now to the Edge Group or CDN financial model. Overall, the business is a more mature business, slower growing than our Security business. The gross margins are lower than what we see in Security, primarily driven by the large traffic-pushing customers in our media and high-tech verticals. You can see that we believe we can maintain gross margins in the high 60s despite some of the challenges around pricing and the near-term headwinds in commerce and travel, driven by our continued focus on server efficiency and lowering the cost to serve traffic. Non-GAAP operating margins in this business are in the high 20%s. That's primarily driven by a lower gross margin in this business. As you can see, this business has a higher capital intensity for network CapEx due to the amount of traffic that we serve.
It's important to point out that despite the lower margin in this business, having this business and having all this traffic enables us to build out our massively distributed Edge and do so very cost effectively. Remember, all the high-margin Security, Edge compute, and acceleration products all run on this same platform. Our R&D expense as a percentage of Edge revenue is roughly half of what we see in the Security business. The R&D spend in this group is primarily to enable the very fast-growing Edge application products, as well as innovations around server efficiency and quality. Quickly bringing this all together, this slide shows the last five-year trends that I just spoke about, and it also outlines the three-five year growth goals for both Security and Edge technology.
If we execute on all the things that we talked about today, this would result in a three-year growth for the company on the top line of roughly 6%-10%. As I move on to the total company margin, this slide shows if the business mix changes over the next three-five years as security grows faster than CDN, the margin for the company should naturally expand, and the capital intensity should also decrease. This slide is an illustration based on the models I just walked you through for each business. You can see here, as we move from just over one-third of our business coming from security with mid-30s net non-GAAP operating margins closer to a 50/50 mix over the next three-five years. You can see our non-GAAP operating margins could expand by as much as 2%.
We could see our CapEx decline by roughly 1%-2% as well. That said, we want to ensure we always are balanced in our approach to expanding margins and investing for growth. M&A could have short-term impact on margins as we scale any acquired business. To summarize, based on our 2021 guidance, which has not changed from our earnings call on the 9th of February, it calls for 5%-7% top-line growth and 30% non-GAAP operating margin. If we execute on the goals that we outlined here for the next three-five years, our target model would result in 6%-10% top-line growth. Gross margin should remain in the high 70%s, and operating margin would be in the range of 30%-32%.
We believe our top line in the security business will come from further penetration of the install base, new product development in M&A, security-led new logo acquisition, and channel leverage to drive our Zero Trust adoption. We expect our Edge growth will come from increasing use cases for Edge computing, especially as 5G and IoT becomes more pervasive, as well as traffic growth from OTT and gaming. On the margin side of the equation, in addition to the mix shift, we will continue to look for margin expansion from operational efficiencies. Now, before I conclude, I just want to spend a minute on our balanced capital deployment. We have a very strong cash business with incredible cash flow generation. Our primary objective when it comes to capital allocation is to leverage our stock buyback program to offset dilution from equity compensation over time.
You can see from the chart, we've been opportunistic over the past few years with our buyback program to reduce shares outstanding from time to time. We will continue to consider this going forward. However, our primary desire is to leverage our cash and cash flow for M&A. You can see from this chart that M&A has been and will continue to be a critical part of our overall growth strategy. We love to find acquisitions targets that are subscale that we can help scale, which will enable us to grow faster. A good example of that is Asavie. We acquired Asavie for less than one quarter worth of free cash flow. We've already accelerated the growth trajectory of the company in the few short months since we've owned it.
As I noted earlier, a key part of what will enable us to achieve our three-five year growth goals will include M&A. We are very active shoppers and disciplined buyers, and we intend to continue that posture going forward. This concludes my prepared remarks. I will now turn the call back over to Tom Barth to start the Q&A section. Thank you.
Thank you, Ed, thank you everyone for joining us today. We now have sort of a modified question- and -answer session with the executives that you have heard from today. A couple of questions have just started coming in, I'll try and direct them and give you a verbatim. The first one, of course, always goes to security. Rick McConnell, if you would be kind enough. It comes from Keith Weiss from Morgan Stanley. One is, how are you calculating the security TAMs, the total addressable markets? Are those industry or company estimates? Also, can you give us a little bit better overview on the competitive environment and network asset access security, and what advantages do we have over more traditional vendors?
Sure. Thanks for the question, Keith. Appreciate it. Relative to the security TAMs, we typically will use Gartner and IDC, most of all, to start with a top-down view of markets and market sizing. We apply our bottom-up propensity to buy based on our customer base and our targeted segments. That's how we come up with the TAMs. We do go through a pretty significant and sophisticated process to evaluate those TAMs as we look forward as best we can as we look to the future. Relative to the competitive environment and network access security, I think the starting point is it's a very nascent space. It is still developing and developing rapidly. There are going to be several leaders in this space, and candidly, we expect to be one of them.
In terms of advantages for Akamai, we obviously leverage a fully integrated portfolio from Ion for web performance, going to Site Defender in web security and web application firewall, as I covered earlier in my prepared remarks. That fully integrated product portfolio obviously extends into our network access security capabilities as well. We have enormous scale and capacity out of our platform that Adam talked about earlier. We have huge amount of visibility and threat data, if you will, from 3 trillion DNS requests a day, multi-billion client devices that we evaluate. That threat visibility is incredibly valuable based on the amount of traffic that we see in the network from Akamai. Finally, we have our Security Operations Center, and that 24 by 7 monitoring provides significant advantage as well.
In summary is we believe that we can compete and compete very effectively in this space.
Great. Thank you, Rick. The next question is to Adam. Adam, this is regarding EdgeWorkers as well as Edge applications. If you could talk a little bit about some strong growth expectations, what are some of the key growth drivers in terms of Edge applications? The follow-on question is, will we support more than JavaScript for EdgeWorkers?
Sure, Tom. Let's start with the growth ones. I covered a few of them in the presentation. The first one really is continued adoption of our EdgeWorkers and EdgeKV platform, especially as workloads migrate to the Edge. Even if you think about the slide that I showed of Gartner's prediction of 75% of workloads migrating from the central clouds to the Edge by 2025, even if they're a little bit off, that's still tremendous amount of workloads that still need to migrate by 2025. The second one is more out-of-the-box solutions like our first-party applications, like Image Manager and Video Manager. The last one is what we're really excited about as well is our app marketplace and third-party apps where third parties will build applications.
We're already seeing demand for all three areas now, and we expect that to continue through 2025. I think, Tom, the first one was about JavaScript. We selected the Chrome V8 runtime engine because it has native support for both JavaScript and WebAssembly, and it also provides an easy path to support other languages. We're primarily seeing demand for JavaScript, which is why we chose that. It's also a language that most developers already know. If we see demand for other languages, we'll absolutely add them to the roadmap and expand.
Great. Well, thank you, Adam. Tom, the next one is for you. It comes from Will Power, from Robert W. Baird. It's with the security growth guide, how much of that is tied to M&A, and how comfortable are we with the M&A pipeline, given such high valuations out there?
Well, that assumes normal levels of M&A that you've seen from us over the past several years. We do look at a lot of deals, we're also very disciplined buyers. As you know, there's some crazy valuations out there, and you're not going to see us do anything crazy. We'll make acquisitions that make financial sense, can bring new products to our platform for the benefit of our customers, and we'll do tech tuck-ins. If we were to do something larger, that could be upside, we're not counting on something dramatic or different than what we've done in the past to achieve our goal.
Great. Also from Will is a question for you, Ed, is how does the shift to a greater security concentration over time impact capital intensity for the business?
Yeah. Thanks for the question, Will. I outlined on one of my slides where I showed what would happen as the mix shift changes. Obviously, today, the biggest portion of the business is coming from the delivery part of the business in CDN, and of that, it's mostly coming from our media customers. There's still a high bit of CapEx in the business. You can see over time as the mix shifts to, say, 50/50, you could see probably a point or two of capital intensity diminish a bit in the business. I tried to highlight for you the R&D CapEx, you really should be thinking about that as R&D spend, not as your traditional CapEx. It's the network line, and as I pointed out, the security business has a much lower network CapEx intensity. Over time, that'll go down.
I'd say probably a point or two as we move towards a 50/50 mix.
Great. Tom, coming back to you. It's from Ghassan Abdo from IDC. He's interested in your view on the competition from the hyperscalers as they continue to accelerate strategic partnerships with mobile 5G.
Sure. We've always competed with the hyperscalers. They're also some of the same ones that we compete with and have for over a decade. They're some of our largest customers. Akamai's capabilities are really unique as we talked about with our edge platform and a fully distributed architecture, and I think that becomes even more relevant in terms of competing as you move towards a 5G model, which will have many points of presence, and Akamai's capability there is really unique in being able to manage that and give great performance. Our underlying software just performs better and is much more secure than what you can get from the hyperscalers, and that's why they are among our largest customers. I think as both Bobby and Adam pointed out, our advantage gets even stronger with 5G. Carriers are very close to Akamai, as Bobby and Adam talked about.
They are big strategic partners for us. We really help them. In some cases, the hyperscalers are more competitive with the large carriers. I think we have an advantage that way as well.
Great. Thanks, Tom. A three-part question now from Jim Fish from Piper Sandler. The first one's going to come to you, Adam, it looks like. How does going after standalone security sales investments as well as targeting the developers actually use EdgeWorkers and KB rather than your competitors? Well, why use Akamai instead of our competitors for EdgeWorkers and KB?
Sure. I guess what I'd say is first, dealing with a lot of customers over the years, in both the services function I ran and of course, dealing with sales over the last few years, most customers really like to standardize on a provider that can really satisfy many use cases at once, right? When you think about Akamai and you think about security, compute, and delivery and doing all that at scale, nobody does that like Akamai. When you go to a customer, being able to provide all those solutions at once is something that a customer really values. We have leading security solutions that are highly differentiated across all the areas that Rick spoke about.
The other part is we've invested really heavily in the DevOps or developer friendliness that I spoke about, and it allows our customers to easily manage across all of those different solutions in a more simplistic way. Again, standardization for that customer across three types of solutions really will make it easy for them.
Okay, great. PJ, this one is for you, is how many reps will be in the Advanced Technology Group over the next one to two years?
Yeah, thanks for the question, Jim. I think the right way to think about the resourcing in that group today is several tens of resources. We'll be careful about our investments moving forward. We do expect to get a fair amount of offload from our carrier partners who are very well-versed in security. They represent several third-party products and have been representing EAA and ETP and some of our enterprise security products for a while. So we'll invest where it makes sense. Right now, we have a really capable selling organization who's been augmented by this advanced technology group for a while, and our sellers are pretty comfortable representing our capabilities successfully in front of customers today.
Tom, this last question is more directed toward either you or Ed. I'll start with you, Tom, which is why not spend the strong operating margin growth or appreciation and go after more opportunities aggressively in security as your competitors are hoping for 20% plus long-term operating margin?
Certainly there's a lot of companies out there that aren't worried about margins or even being profitable. At Akamai, we really do care about profitable revenue growth. Security, we're making the investments we think we need to maintain a 20% growth rate. We'll continue to do that. It's not impossible from time to time as we make larger investments in an area or do an acquisition that it could for some relatively short period of time, depress margins. That's quite possible. We do believe that we can operate with strong operating margins while we grow the business. I think that's important for the long term. I think we're in a phase now in the market where you see companies that don't care about profitability. That's not the case at Akamai.
Okay. Rick, coming back to you on security. On the fourth quarter call, it sounded like Asavie had made a really strong start under Akamai. Could you provide more color on the integration since then, and if there are any updates to the guidance on the $30 million run rate?
Sure. Integration has gone exceptionally well. We're very pleased with how Asavie has integrated so rapidly into Akamai. No change in guidance at this stage relative to our call that we just completed recently, but we're very pleased with how it's gotten out the chute and how it's integrated so rapidly into our access control portfolio.
Okay, great. One follow-on from Rishi over to you, PJ or Adam. I'll start with you, PJ, which is, we've seen some comments on pricing pressure from peers in the space, and he wanted to get a sense, is there any industry-wide issue or something else that Akamai or you and your teams are seeing?
Yeah, I don't think we've seen anything unusual in terms of the competitive environment. Volume drives price, especially in media. I think we've been able to do really balanced deals with a lot of our large OTT customers. We haven't seen anything unusual in terms of pricing, especially in media. It's pretty standard, and we're obviously working hard to drive costs out of our business so we can pass those savings on to our customers.
Okay. Adam, I'm going to come to you. The question is around SaaS platforms. As they grow as a percent of e-commerce, how important is it for Akamai to establish relationships with Shopify, Salesforce, and other software companies to their end merchants and customers?
Yeah. We view providing services to platforms as a service or any type of SaaS platform as critical. Over the years, we've been already doing that. When you think about what we do with our partnership with Microsoft and being available in their marketplace, we have a solution that we've worked on with The Washington Post, with their Arc solution, where we already provide solutions out to their publisher customers. We are targeting to work with all SaaS platforms like that because we do believe it's important that Akamai security and delivery becomes available to their end customers, whether they're buying it directly from Akamai or buying it through other providers like this. I don't know, PJ, if you want to add anything.
No, I think you're right. I think we're going to continue to innovate there. Obviously, there's I think, a big opportunity for us to do more with a lot of the SaaS PaaS providers, but I think we have a pretty good history of bolting on our capabilities to the Azures and the Rackspace and the IBM Cloud. We have some experience there, and I think we'll continue to augment what we have in the market, where we think it makes sense to do so that we can tap into new opportunities.
Okay, Bobby, I'm going to come to you, if you don't mind, from Sterling Auty, from JP Morgan. Could you talk a little bit more about how you're planning to leverage edge capabilities to capitalize on the 5G opportunity?
Sure. Thank you. One thing, of course, as we've talked about 5G, of course, makes the edge more and more important over time, and I think whatever comes after 5G, presumably 6G or whatever, will only serve to continue that trend. Second, I think it means that the mobile carriers play a more and more important role in the ecosystem, overall ecosystem. Our partnerships, which we've had these strong partnerships for a long time, we have more partnerships now as we've added Asavie to our portfolio. Those become more valuable. Our go-to-market motion that PJ talked about, working with our carrier partners becomes more important. I think basically it is that center of mass moving to the carriers, to the edge, and our partnerships that I would call out here.
Okay, Bobby, I'm going to stay with you from Brad Zelnick from CS First Boston. You mentioned a number of times that delivery and security run on the same network. How many of your points of presence run on the entire set of solutions that you have in the market?
Well, I don't have the number, but when we talk about locations, countries, locations, networks, and things like that, the vast majority of our functionality runs across all of those locations.
Okay. The next question's from Rob Majek from Raymond James. It goes to you, Rick, if you don't mind. On security access control, you mentioned a doubling of the revenue to approximately $100 million this year. He believes this includes inorganic benefit from Asavie, which would imply organic revenue growth of roughly 40% in 2021. Can you help walk us through how you think about organic growth rate for the access control business in the next few years, given the large opportunity investment in EAA and ETP?
Sure. Thanks, Rob. Let me break it down into 2021 and then maybe look further into the future from there. For 2021, the organic growth rate, based on the numbers that we provided, is actually north of around 60%. That's the organic piece. The overall $100 million does include Asavie, so it does include the balance of that being inorganic. As we look to the future, we talked about access control growing at north of 50% over the three-five year time horizon as a goal, and that's what we're after, and we believe that's going to come through substantial incremental organic growth in the future, as well as it being an opportunity for further inorganic expansion as we look over this time horizon as well.
Rick, I'm going to stay with you for a follow-on from Rob. You have a huge opportunity in access control between EAA, ETP, MFA, but you also have a large number of well-funded competitors. You walked through some of the key differentiators in Akamai to win business, but it would be great if you could elaborate further in terms of how do we compete in that space.
Okay. Well, I point back to my earlier commentary on how we differentiate. The other thing I'd say is that, sure, there's some large, well-funded competitors, but we're large and well-funded also. We're investing in these areas aggressively, and we believe that this is a substantial business opportunity for Akamai as we look to the future. Very excited about the multi-factor authentication launch that's forthcoming here, as well as what we're doing in EAA, ETP, and of course, the balance of the access control portfolio. This is an area of strong investment for us. We're going to continue to invest in it, and we expect that it's an area where we have substantial differentiators, as we talked about earlier, and that we can win in the market.
Tom, I'm coming to you for the next one from Brad Zelnick from CS First Boston. With the guidance for 18%-20% security growth in 2021 and an expectation of 20%+ through 2025, should we expect that a mix shift is driving that growth, or should we expect to see greater velocity of M&A near term to help drive higher levels of growth?
Good question. We do have some very exciting areas, for example, network access, which Rick talked about, that are growing very rapidly on smaller numbers. As they become larger, that does help accelerate growth of the security business overall. That said, as we talked about, we're getting very strong growth pretty much across the board in our security products. We don't have plans, I'm certainly not relying on some big acquisitions to attain the goals, as I talked about. As I mentioned, we've guided to 18%-20%. Our goal is to do as well as we can, and we'd hope to do better than that. That would, of course, be consistent with maintaining a 20% + growth rate.
Ed, we're going to stay on M&A. The next question is about your comment on disciplined M&A strategy. Can you provide a bit more specifics, i.e., is there a long-term hurdle on return of ROC that you're targeting, and how have your acquisition performed?
Yeah, great question. We're obviously looking at whenever we go into a product area, you look at build, buy, partner, and you do the analysis and determine what's the fastest way you can get to market and get the quickest return. We look at M&A as an augmentation strategy of our overall products and growth strategy. Not a specific ROIC that I would throw out on the call, but we do look for things that can run on the platform, look for assets that are subscale that we can scale with our go-to-market and our relationships. I talked about Asavie as being a perfect example. The big challenge for them as an independent company was getting mind share within the large telcos. We obviously have that. They also need to build out their network presence as they add these telcos. We already have that.
That's a perfect example of something that we could have built on our own. It would have taken us a little bit longer. These guys were at a stage where they were looking to scale. Perfect example of how we tuck in M&A into our strategy.
PJ, coming to you next. You talked a little bit about it in your presentation, but can you talk about how Akamai will attract more traditional security channel partners as EAA and ETP become larger demand items?
I think we'll be measured in terms of how many new partners we sign up. Certainly, if we have aligned goals and interests around new business generation or exploring new emerging markets together, we'll sign up more partners when we need to. Keep in mind that 60%+ of our new customer bookings came from channel partners in 2020. We have a really healthy and invested partner ecosystem today, and I think you're going to see us align more closely with our carrier partners who have access to enterprise accounts that we want to have access to. I think they're gravitating towards our market-leading capabilities. I think there's more to do with our existing partner ecosystem versus all of a sudden expanding that partner ecosystem to layer in a bunch of new partners.
We'll be careful and measured in terms of the new kinds of partners we sign up.
Thank you, PJ. I think it comes to you, Bobby, but if you'd like to ask Rick to help or whatever, it comes from Alex Henderson from Needham. It is around Enterprise Application Access. How does Akamai deliver policy? Is it on a per-app basis? Is it on a per-user basis, or is it based on traffic flows?
Yeah, it's a good question. When you think about access, and you're making a decision on whether or not to grant access, it's based on a number of factors. One factor is who is the user. You need to connect with some kind of identity capability so you know who that user is. Second, what is the device that they're using, and what is the profile, of course, of that device, the operating system, whether it's a managed device, whether it has antivirus on it, things like that. Context, where are they coming from, what country are they connecting from, things like that. Then what application are they connecting to? Those four factors come together to make a policy decision on whether or not then to grant access.
A key point, by the way, to add is that in our case, once we grant access, we stay in the traffic flow. Many solutions for access will grant or deny access, but then they get out of the traffic flow. We stay in the traffic flow, and the reason is because then we can inspect the traffic flow to make sure there is nothing bad in that traffic flow. For example, apply web app firewall rules, and not to mention accelerate that traffic flow. Yeah, probably a longer answer than you were looking for, but there it is.
Okay, great. Thank you. The next question, I think, Ed, I'll come to you, it's from Brandon Nispel from KeyBanc. What is the company's philosophy on capital allocation? What is the leverage target we should look for? What level of buyback? Why is the company not currently paying a dividend given the stability of revenue growth, margin expansion opportunity, and ultimately free cash flow generation?
Yeah, great question. It's something that we talk about all the time. We've just created a subcommittee in our Board of Finance Committee where we talk about these things on a very regular basis and always think about what is the overall total return to shareholders. Right now, we're in a position where we believe there's a lot of strong growth ahead of us. Using our capital for investing back in the business and M&A is the primary area that we want to invest our capital. As far as leverage ratios, think of us as sort of an investment-grade level company and maintaining leverage ratios that are in line with that, I think is very smart to do. Obviously, if we find larger acquisitions, we could tap the capital markets and raise some debt to finance a larger acquisition. That hasn't been our strategy so far.
As I talked about with share buybacks, we've been opportunistic over time. We always talk about looking to do that when it's appropriate. Primary use would be to offset dilution and then for M&A, like I talked about.
Okay. Adam, I'm going to come to you. It comes from Jim Breen from William Blair. Can you discuss the longer-term views post-COVID on CDN traffic as it is impacted by the OTT providers?
Sure. Thanks, Jim. I probably think about it in two vectors for CDN traffic, both gaming and OTT. Specifically on OTT, I think what we see is COVID itself kind of set a new baseline last year, and it accelerated adoption of these online services. It happened to also coincide with the launch of a lot of new online services. Kind of had two things happening at the same time that really accelerated growth there. We believe it's going to continue out into the, let's call it, near future to foreseeable future as those services become more globally available. They started in the U.S., as they expand to Europe, Asia, of course, Latin America. I think what you'll see is continued similar growth that we have seen over the years of CDN traffic growing with both OTT providers expanding.
Like I said, I think it is important to think about gaming and the expansion of free games and the games moving to digital updates and more of the expectation around constant updates to in-game content, right. What you see with things like Call of Duty and Fortnite offering up constant updates to their games to provide a more living environment for their users.
Rick, I'm going to come to you next. It's from Colby Synesael from Cowen. Could you help us unpack what is driving lower security growth in 2021 versus your three-to-five-year security CAGR? Also, could you help give us a sense for the M&A priorities within the security business?
Sure. Thanks, Colby. The first element is just to come back to both what Tom and I mentioned earlier, which is our guidance is at 18%-20% for the security growth during 2021. Our goal is 20%+, that's what we're going to strive to achieve and deliver. In terms of the longer-term CAGR, obviously M&A and inorganic is a component of that growth and maintaining the growth rates at north of 20% in that CAGR goal. Furthermore, we've got a number of product launches coming up. We've got the multi-factor authentication. We've got Account Protector. We continue to invest organically in new product lines and new product areas, all of that gets factored into our view of the long-term CAGR.
In terms of the M&A priorities, you could see investment from us through inorganic investment really in any of the four categories that we've indicated here in security. At least at the moment, we're principally focused in the areas of fraud prevention as well as access control, which is where you're going to see the highest growth as we covered earlier.
Ed, I'm going to come to you. Coming again from Keith Weiss from Morgan Stanley. Why doesn't gross margin improve in the target model if you have a much higher mix of security revenues?
Yeah, good question, Keith. I tried to outline the different components inside of each one of the businesses. If you look at CDN, for example, which is the largest part of our business, the delivery side, there's even components within there. I talked about how if the growth is being driven by media traffic like we're seeing today, that's at a lower gross margin than what you see with application acceleration mix. You'd need a bit of a mix shift there as well. I think over time, if I were to take out that chart that I showed at a 50/50 mix to, say, 2/3 or 3/4 from security, you could potentially see some increase in gross margins as well.
For now, as I model it out, I'm still assuming that we're going to have a pretty healthy growth on the delivery side coming from media, which will be at that lower gross margin. It's really just a balance.
Okay, Adam, coming to you from Rob Majek from Raymond James. You mentioned Edge delivery revenue likely declining this year from a tough year-over-year comps. Can you elaborate further on that point and comment on the potential extent of decline and help us understand how to think about Edge application revenue growth in 2021 that may offset some of that?
Yeah, we're really excited about the Edge application growth rate as it accelerates, particularly into the future. It'll really help the overall CDN revenue number, though, as that number gets larger as a total percentage of the CDN revenues, I think as you think about out of 2021. I do think, as I started before, I think you really just should continue to think of the business as somewhat flat to upwards of 3% growth over the total time horizon. Really, my comments were really specifically geared around just re-emphasizing the potential uncertainty of the future of COVID and particularly the future of those specific industries like travel and commerce in light of COVID.
Okay. Sorry about the cat, folks. It's live show, right? You never know what's going to happen in a live show, right? At least it's not the cat filter. Kim, I'm going to come to you on how do you increase awareness of Akamai as a premier security brand?
Sure thing. first
Kim, you may have to unmute. Okay, great.
focus on building awareness. Okay. I said, first we focus on building awareness in the places our customers and prospects rely most, the top three being analyst reports, tech websites, recommendations from people inside their org. Of course, we focus on initiatives to amplify our brand and awareness. Anything from increased analyst engagements, PR, case studies, working with CXOs and developers. As I said earlier, basically, we focus on driving awareness with the people and in the areas that matter most to us.
Okay. Bobby, from an investor, comparing centralized versus decentralized network. With advances over the last 20 years, can a centralized model work better now than it may have had in the past?
Good, thanks for the question. Well, the answer, in short, is actually just the opposite. Let me explain. To begin, I think it's worth recalling that the Internet performs very well and scales very well in very large part because we're at the edge and have taken so much traffic, continue to take so much traffic off of the peering points, off of the backbones and so on. Second of all, as you look at the changes that you're talking about, how has the Internet changed? The big changes that you really want to point to are things like the advent of 5G and broadband technology that adds capacity, huge amounts of capacity at the edge, and increases demand at the edge. I would add on IoT. Think about all these devices that are connecting and the amount of data that they're putting through the network.
Think about big data and AI, massive amounts of data, cybercrime. All of these things are actually shifting the center of mass of the Internet actually toward the edge, not away. What we've seen over the last 20 years is that it's become steadily, clearly harder and harder to deliver traffic across one network to another through peering points, harder and harder to deliver from one city to another across backbones. Increasingly, it's important that you avoid those things. You have to deliver from the edge. The center of gravity moves toward the edge. If you look forward, you just have to look at the trends and ask yourself whether or not these are durable trends or whether these are here today, gone tomorrow trends, 5G and whatever comes next. IoT, big data and AI. I mentioned cybercrime.
If you believe those things are going away tomorrow, well, you might reach a different conclusion, but I don't think anybody believes those things are going away tomorrow. Those trends are here with us for a long time, and they're only going to serve to make the edge increasingly important. Everything that we've seen in the Internet has actually moved toward an edge model and away from a centralized model, not the opposite.
Okay, Tom, I'm going to come to you. It's a question from an investor about how engaged is our board, particularly the new members, and do they look at some of the things that we've done in the past or future acquisitions?
Yeah, we have a very active and engaged board, all the directors. We do, as a board, review the success of our acquisitions on pretty much a quarterly basis to track progress and are we meeting the objectives of the acquisition.
Okay. Moving to you, Adam. How concerned should we be about potential for significant DIY headwinds?
Yeah. This is one we get a lot, and I think the way I always like to think about it is, while you saw the giants that we report on do some migration of traffic or content onto DIY services, there's still a significant role for a provider like Akamai in a world like that. In the case of OTT platforms, they're maniacally focused on quality. They need tremendous number of redundancies to ensure that kind of quality for their end users, which also means that they're not going to go to a single platform. I think the other piece here I'd say is that we have a very diverse customer base, so even if one customer were to move some traffic to a DIY platform, you're not going to see a significant impact in terms of revenue on the Akamai platform.
There will be people that use DIY, but again, I think we've shown track record, particularly with those giants, where there's still a tremendous role for third-party providers.
Okay, great. Adam, I am going to stay with you. It is from Alex Henderson from Needham. Can you talk about how many coders are using your Workers tech?
Great question. I don't know the number of coders per se. What I'd say, probably the way that I'd tell you the answer to this is really thinking about the amount of requests being handled on the platform. We've seen, I think it's something like a 9x quarter-over-quarter growth of requests on the platform. I think it was three billion requests handled in Q4, might have been in December, actually, but on the platform. We're performing at scale. There's a lot of customers using this, but I don't know the exact number of coders. I don't know, Tom.
About 300 billion requests.
Oh, did I say $300 million? Sorry, I meant $300 billion.
Yeah.
Sorry, Tom. Yep.
That's great. All right, a couple from James Fish. Bobby, I'm going to start with you. I'm pretty sure it comes to you, but maybe Rick. On the security side, how do we think about providing the SD-WAN connectivity side to create a full SASE?
Our focus in SASE or Zero Trust or in general in access is really on the security. The goal here is to be able to protect all enterprise traffic flows no matter what they are. For the network part of it, the best path forward is really to partner with our carrier partners and other technology vendors, to provide the actual connectivity component. We'll focus on the security part. I don't know if you wanted to add anything, Rick, on that.
No, I think that's right. Generally speaking, as we talked about in our earlier remarks, I think that we are undergoing imminently a significant shift in network topology. As that shift in network topology happens, I think it's going to move to a different framework, and it's the framework of a Zero Trust Network Access kind of provision. I think that's going to impact how customers and companies think about everything from SD-WAN to other forms of connectivity. As Bobby indicated, we're focused on the security aspects of that primarily.
Tom, I'm going to come back to you. Sorry. I'm going to come back to you, with a question regarding blockchain. Could you provide us just a quick update on the Mitsubishi partnership and anything you might think about that future opportunity?
Sure. The partnership is called GoNet, and Akamai has about a 20% share of that, and we provide blockchain services to GoNet. We built the world's leading blockchain platform in terms of volume, with how many transactions can be done per second, PCI compliant, and very low cost per transaction, which is a big improvement over what you see out otherwise in the marketplace. GoNet itself has been a little bit slower going to market in terms of its revenue generation than the initial plan. We are hopeful for the future that it will be successful. We didn't call it out today per se, because it's not a big source of revenue for us now. That would be upside in the future.
Thank you, Tom. PJ, I'm going to come to you. Prices are continuously declining in CDN. How is the pricing flexibility and how do customers influence the price?
Yeah, good question. I think more than anything else, volume influences price. As I mentioned before, we've been able to do really balanced deals with a lot of our high volume OTT and gaming customers. Some of these customers have tiered pricing in place, so really the risk is balanced. We're not taking on too much risk, and neither are they. I think what we found is in a lot of the big high volume providers really value scale and reach and performance. As these business models become more mature, especially in OTT, they're paying very close attention to QoE. They're measuring lots of things related to performance. I think there's certainly advantages of being on the Akamai platform, in terms of our reach and the kind of performance we can give a lot of these OTT providers.
We're able to do really balanced deals, and we're pretty happy with the outcomes that we've seen in the OTT and high volume media space.
Tom, from an investor coming back to you regarding the restructuring and realignment, why now, and what are the immediate and long-term benefits that you expect?
Yeah, good question. The short answer is it makes sense to do. Bringing all of our security businesses together into one unit just focused on security, I think is a good thing for us to do. We're now at $1 billion. Our enterprise unit, which was in incubation mode for a while, and our carrier security unit both are reaching critical mass. Now that together with network access, we hope will be $100 million this year. We said we'd be calling it out to the street when it got to $100 million, and I'm hopeful that'll happen this year. Bringing together the teams that work on platform and content delivery, I think makes a lot of sense. I think we'll operate more efficiently and be more agile. Bringing together the sales force makes sense now.
When we started with a split sales force five years ago, that made sense because media customers bought media products and web customers bought web products. That's before we really had much of a security business. Now at $1 billion, all of our customers are buying security, and our media customers are some of our biggest security customers. We're in a different situation today than we were five years ago, actually much stronger as a company. It made sense to do. At Akamai, when there's something that makes sense to do, we generally don't wait around. We just execute on it.
Bobby, coming to you from an investor. Some of your competitors have talked about having more efficient architecture and networks. I know you've covered a little bit, Bobby, but in particular around Akamai having multiple networks and/or older hardware. How would you respond to this?
I think one is we're always innovating and making our network more and more efficient, and I think you can see the results in our numbers, which I think speak for themselves in terms of our efficiency. In terms of the hardware, on any given year, as we acquire new servers, we're always using the latest technology that makes sense for our workload. We work very closely with the hardware providers, and even do some amount of customization in various aspects. We are, year over year, always using the latest hardware and most efficient hardware that makes sense for our workload. We're also innovating on the software side to make sure that our software is as efficient as possible, and I think the numbers speak for themselves on that.
Okay, Ed, I'm going to come back to you from another investor, and I'm just going to read it. It said you did an admirable job sidestepping the dividend question. Asking it more specifically, why has Akamai chosen an opportunistic share repurchase program over a dividend, either regular or special? In what conditions would change the calculus around a paying dividend?
Yeah, I did not mean to sidestep the question, so I apologize for that. Obviously, as you look at your total return in capital, you look at everything. I think our investor base, you can tell by a lot of the questions that we've heard today, are more interested in growth. If you look at, say, for example, the average S&P dividend yield today is about just under 2%. I think our investors are looking for a greater return than that. That would be about, call it, 30%ish of our free cash flow, and we think we can get a better return by either opportunistically buying back some stock or really putting it into M&A. That's where you get the best return. In terms of how we think about overall capital, as I talked about, it's something that we think about.
I don't think we're at the point now where our investor base that we have today would like to see a dividend. I think they'd rather see us go for a higher growth rate. It's something we'll consider in the future for sure.
That's an admirable re-answer of the question there, Ed. Good job. I'm going to move on to, it looks either Rick or Ed, but from Colby Synesael from Cowen again. Can you help us separate the organic growth embedded in your security growth CAGR from the inorganic growth contribution?
I talked a little bit about it earlier, the organic growth rate that we see during 2021, as part of the numbers that we provided in access control, is 60%+. That's the organic portion. The balance of the growth rate comes from the inorganic element. As we look to the future, as Tom and Ed both mentioned, we're assuming a consistent level of M&A investment to achieve those growth targets over the three-to-five year horizon. If we see something that's larger or more transformational, that could provide some upside.
Rick, I'm going to stay with you, which is from Alex Henderson from Needham. Why would you launch MFA when Okta and others do identity so well?
It's a fair question in terms of MFA. We look at participants in the market like Duo as an example, and candidly, we see that it's an opportunity in a specific area that's ripe for some disruption. We believe that we can do it very well with significant security and security improvements. That's an area that is an extension of our access control portfolio and extension of our enterprise access and enterprise threat protection capabilities. That's a very logical and natural evolution for us to provide to customers.
Okay. Well, at this time, we don't have any further questions in the queue. I just want to remind everyone that the presentations are available on the akamai.com section in the investor relations section. Tom, if you wouldn't mind, I'll pass it back to you to wrap things up.
Sure. Thanks, Tom, and thank you all for joining us today. We appreciate your interest in Akamai and look forward to speaking with you at conferences coming up and then various callbacks. Thanks again.