All right. Thank you, everyone, for joining us. My name is Keith Weiss. I run the U.S. software research franchise here at Morgan Stanley, and I'm very pleased to have with us CEO and Co-Founder Tom Leighton from Akamai Technologies. Tom, thank you so much for joining us.
Pleasure.
So I wanted to start off with a big picture question, because I've been covering Akamai for a long time, and the evolution of the story has been pretty fascinating. It's expanded a lot from where even I started looking at it, and hugely from where you guys started the company. But it all has a common core, if you will, and it's the Connected Cloud. You guys have grown this network, 4,100 points of presence, over a petabyte of traffic capacity as of right now. Can you talk to us about to what degree is the strategy? How can we expand? How can we utilize this resource more and more on a go-forward basis? And talk to us a little bit about the most recent kind of big expansion with Linode into a broader computing capacity.
Sure. Obviously, use the platform for delivery. That's where we started by being in 4,000 points of presence. You're very efficient, very low cost, great performance, great scalability, and still true today. Then added security starting a little over 10 years ago. The best place to stop the attack is where the attacking entities start, so intercepting that traffic and it's all done as part of delivering the traffic. Same CPU, same communication sequence. Before we deliver your bank balance, we're checking it's really you, and not some bot or somebody that stole your credentials, for example and then most recently with compute, actually getting the compute instance close to the end user has those same advantages. It gives you better performance and lower cost, better scalability, ultimately better ease of use.
That platform is a common base for a lot of what we do and very compelling advantages for our customers.
Got it. Got it. So I want to start digging into the Compute side of the equation. It's the fastest growing part of Akamai right now, and it's one that I think has captured investor imagination on kind of where Akamai could go on a go-forward basis. There was an evolution, if you will. There were Compute capabilities before you guys acquired Linode. There was the ability to do some programmability on the edge. Can you talk to us about bringing Linode into the equation? What were you guys able to do before, and then sort of where did you get to with Linode? What was the buy equation, if you will, or the rationale for making that acquisition?
Yeah. So before Linode, we would support JavaScript on the edge. We'd spin up your JavaScript applications in about five milliseconds based on end user demand, so very fast, very scalable, easy to use, but it wasn't supporting your containers and VMs, which the vast majority of the compute marketplace, that's what they want, the customers want. And Linode gave us that capability. Now, since then, we've done a lot of upgrading on capabilities and capacity for Linode. Linode originally was targeted at what small and medium businesses need, what developers need, very popular with developers, but it wasn't at a level for mission-critical apps for the world's major enterprises. And that's what we're in the process of doing and have made a lot of progress on.
Yeah. I was going to ask you about that, because a lot of investors, including myself, were a little bit surprised, because the core Akamai customer tends to be a larger enterprise or a larger organization. Linode was very successful with small companies, and ease of use was a big component there. What did you see in Linode that gave you the confidence that you're going to be able to take that small business-oriented platform and make it accessible, make it usable by your large enterprise customers?
Because that's our skill set. That's what we know how to do. We know how to take things and make them suitable for the big enterprises to provide the scalability, the reliability, the security, the certifications that major enterprises need and trust Akamai to do, and Linode had the basic functionality, very easy to use, the container support, VM support, which we didn't have that, and it is important to us that developers really like it and find it easy to use, because ultimately they're the ones writing the applications, and the fact that today we can go into a major enterprise and talk about Compute, and generally there's a Linode user there that really likes it, that's very helpful for us.
Yeah. So the developer side of the equation is critically important to any platform, particularly one where you're looking to get your customers to develop new applications for. It's something that the hyperscalers have always focused on a lot. What's the strategy for driving that developer affinity towards the platform? How do you assure that? Because developers can be finicky. Developers hop around to the best tools. They're able to utilize different technologies pretty easily. What's the strategy for ensuring that Akamai and the Akamai computing platform more broadly maintains that momentum that Linode brought in the door?
Yeah, so you keep it easy to use. You enable the platform with the functionality they need, and we add the capabilities the enterprise requires. I think the situation had been that, yeah, we go to a meeting and there's a person there that uses Linode, likes it, but they're not using it for their big applications, and that's what's changing now. Yesterday we had a lunch with a lot of the major media executives here in London, and it was a big gaming company, and he's now a senior executive in charge of all their technology and development, and sure enough, he was a Linode user and loved it, and now they're migrating their mission-critical stuff onto Linode, which he couldn't do before the acquisition, but now is happy to do.
Right. So that motion of making Linode more enterprise-ready, more mission-critical, enabled, more secure, does that start to put you into more direct competition with the hyperscalers? Because I know when you first made the acquisition, you said, hey, listen, there's plenty to go around. We're going to go after the workloads. Perhaps the hyperscalers aren't. Does this enable you to maybe more directly compete?
Oh, yeah. No, we're definitely competing with the hyperscalers. Now, that said, we've been competing with them for a long time in delivery and security. Two of the big three in the States are actually large Akamai customers. We compete with them, but they use Akamai for delivery and security. And I think over time they're going to use us for Compute as well, because not all of their Compute, but Compute where they really care about performance and data moving around that they're using Compute for end users so they want to be close and caring about cost. We're very competitive there.
Got it. Can we talk about what workloads you're seeing on that edge? There's some components that I think investors have a good idea about, like geo-specificity of applications. That makes a ton of sense, because you have the 4,100 points of presence. You can get very close to that edge user. What are some of the emerging applications that take advantage of the super low latency that you're able to provide or that distributed nature that perhaps investors aren't thinking about that could be that big opportunity over time?
Yeah. Applications that are being accessed and used by end users is really a sweet spot, because often latency matters. So you're buying something online. You want that experience to be fast. And there's a lot of applications involved in that, including deciding what to show you based on what you've been doing so far, inferencing, ad selection. What ad should you see based on what they learn about you and what they see happening? Even video, tailoring that experience based on the end user and the device live and on demand. Anything so there's an end user involved and there's money associated with it for our customer, because a better, faster experience, higher quality translates into better business for our customer. Then there's also the cost side, especially if there's a lot of data moving around, which we're really good at and really efficient at.
We have much lower cost than the hyperscalers charge. So if there's apps that are like that, of course, media is very intensive that way. Commerce, high tech tend to have a lot of data flow. So that gives us an advantage over the hyperscalers for our customer.
Got it. Got it. So we made it 10 minutes without mentioning generative AI, but that's as long as I can go. Can you talk to us about to what degree do you see the Compute platform as a beneficiary of generative AI, either from a revenue opportunity of specifically you mentioned inference before, specifically well-suited for those types of workloads, but also how can you use it internally? Because you guys are always talking about efficiencies, making the entire platform more efficient and bringing down the overall cost.
Yeah. Internally, I think we have over 40 projects now, just every sector of the company at least experimenting with GenAI to see can it help improve efficiency, give a better result for what you're doing, NOC, SOC kind of capabilities, faster diagnosis of what might be happening, a lot of interesting use cases, projects even to accelerate metadata so our customers can more quickly make changes to their web properties and how they perform on our platform. So all sorts of use cases. Now, it's already in two products. One's the Akamai Guardicore Platform as the interface, which gives you a much better human-oriented interface. So a non-expert, say a compliance officer, can interact with what's happening in the internal network infrastructure. They can say, what's that device? Are my firewall rules up to date? How far out of date are they? And get responses.
It can also now suggest initial rules. One of the things you've got to do, like you install Guardicore, and it gives you tremendous power to see what's going on, but also to proactively block a lot of the communications that could be damaging. You've got to set it up initially. So now we can use GenAI to suggest, hey, here's how we think your initial configuration should be, which makes integration much, much easier. We're also now using it with our WAF platform. Again, to help with all this data, what does it mean for you and what's happening real time? I think, yeah, a lot of use cases that'll benefit our products, but also internal operations at the company.
Got it. So the two GenAI sort of applications that you suggested both on the security side of the equation, Guardicore being the microsegmentation solution and the web application firewall, is this starting to suggest synergies showing up between the various businesses of being able to either kind of cross-sell or leverage sort of the solutions across the dividing lines of the segments that we're talking about?
Yeah. So we are selling a lot more of the Guardicore Platform into the existing base. We still have a lot of room to go there. At the very highest level, well, Guardicore is like a firewall. Our agent is a mini firewall that sits on internal applications or devices. And of course, WAF is a firewall that protects you from the outside world and those interactions. And API security is a different kind of, really at a high level, a firewall in a sense, watching everything that's happening and blocking things that are known or thought to be malicious. And so I do think over time there is stronger synergy in those products. Now, today there's different buyers for the WAF platform and the Guardicore Platform. Yeah, over time, good cross-sell, but over time probably there is convergence there.
Got it. Got it. So the cloud platform has been definitely the fastest growing part of Akamai over the past couple of quarters, if I'm not mistaken, mid-20s type growth. How do you guys see that growth profile on a go forward basis? How durable should that growth be as investors look forward into 2025 and beyond?
Yeah. So there's the big cloud number, which includes our older solutions like Cloudlets, Image & Video Manager, Compute that we embedded into our platform for special purpose applications that a lot of customers buy. And then there's Edge Workers, which is JavaScript. The really explosive area for growth is the enterprise, major enterprise use of what was the Linode platform that now has really been upgraded. And we talked about exiting this year at $100 million ARR compared to last year, well, close to zero, because we weren't ready really to sell it. And that's a number that we want to make billions. And the rest of Compute is going to do very nicely, but it's that $100 million ARR that is really the explosive component, because there's just a vast market opportunity there. And early on, the sweet spot of sort of the center of mass is big media.
They hit all the right things. They're big Akamai customers already. They know us. They love us. They use us for delivery and security. They hate writing that giant check to their arch enemy. And we can give a great, better performance at a much lower price point for them.
Amazing. All right. So let's transition to security. Security has been another really expansive business for Akamai over the past decade. Started in DDoS protection, web application firewalls, stuff that was really close to your core user, like the person who was setting up the website, looking to do sort of delivery of the traffic. Now you are protecting that website from denial of service attacks. Now you're protecting that application with a web application firewall. But you've expanded well beyond that. So can you talk to us about the expansion of security, why that made sense for Akamai to provide stuff like Guardicore, like the microsegmentation, the network access controls? Why is Akamai a natural fit for those types of businesses?
Yeah. So first there's a stack on top of web app firewall. So we created the bot management sector. We're the leader there. It's grown very nicely for us. Account protection on top of that. So a human has stolen your credentials. They're trying to access your bank account. They have your credentials. And we detect that and block it. Brand protection for the zillions of phishing sites that have been set up. And that's getting so problematic now with GenAI, because it's so much easier to get people to go to the wrong place today. Client protection so that our customers' clients are protected from malware that's in their digital supply chain that corrupts the client and steals their data. Magecart-like attacks and have leading capabilities there. And next year that's part of a requirement for PCI compliance. You're going to have to have a solution for that.
And then there's the Guardicore Platform, which is today a different stack. That's really the last line of defense inside the enterprise to protect the internal applications that are not necessarily customer-facing. And malware is getting in the enterprise today. And you've got to protect it. You can't just assume you're safe inside your moat because you're not. And so you need to, at a high level, similar kind of protection inside that you would always have now with web app firewall for outside. And so it's protecting the internal applications, detecting when malware's in, blocking it from spreading so that the malware can't move from something internal to a sensitive application or data store. So at a high level, it conceptually is a similar kind of a thing. The details are a little bit different. It lives inside instead of outside facing. Today, the lower level buyer is different.
The CISO is the same. And I think over time there's increased convergence between them.
So how successful has Akamai been in getting the conversation to go up to that CISO level to be talking about a broader security and being able to sell that broader platform, that broader security platform into the end customer?
I think today we're always at the CISO level, any major enterprise.
Now, is there any particular we've heard from other security vendors, and not even just security vendors, like consolidation of spend with fewer vendors, particularly in areas like security, but I think broadly within software, has really been on the rise. We see it in our surveys, and we're seeing companies lean into it. Like Nikesh from Palo Alto talks about platformization. CrowdStrike is doing it with their CCCP program. Salesforce has a Foundations program where there's basically like a freemium program for existing customers. Any programmatic ways that you guys are trying to lean into that consolidation effort to try to get more bundling of the solution or more consolidation of the solution?
Yeah. So we do that in our contracts. We have the Guardicore Platform we launched last year, combines North-S outh and East-W est protections. And we have the WAF platform with a full stack of capabilities built on top. And we write contracts that include both platforms. But I think we're very synergistic with that desire. That said, the big enterprises have dozens of security vendors. And a big bank will have over 100. Hard to imagine, but they do. There is a desire to consolidate, but also a desire to have the best of breed for the key things. And so I think we're in a good sweet spot there with our best of breed microsegmentation as the cornerstone for the Guardicore Platform and best of breed WAF and bot management for the WAF platform.
Got it. So we've talked to Guardicore a couple of times. I'm going to call it your favorite child since how much you've talked to it. We've talked about a run rate of over $180 million. Where do you see that solution going? How big is the market opportunity for this microsegmentation? How big of a business do you think that could become?
I think large. Today you see the regulated industry's critical infrastructure now required to have a segmentation solution. I think long term every major enterprise needs it. Think about it. It's your last line of defense. Enterprises are getting penetrated, and that's only going to increase with GenAI, because it enables the attackers to have much more power to break in. You train your malware around the defenses to break in, and just one open window and you're there, and so you really need defenses inside the moat to detect you've been penetrated, to stop the spread proactively, and that's exactly what segmentation does, but it does it in software now. The traditional approach in hardware was a disaster, which is why nobody effectively did it, but now there's the capability to protect yourself from ransomware and data exfiltration, so I think long run every major enterprise needs that.
We'll have some sort of solution like that.
Guardicore has been a real success for you guys, has grown really quickly to scale. You have multiple scale businesses within that security line. Can you remind me what are the core kind of scale businesses that have kind of reached that $100 million mark within that security portfolio?
Yeah. So I guess there's three pillars, if you will. The one we haven't talked about is the scrubbing or the routing layer denial of service. So just bulk traffic based on without even looking inside the packet. And that's the oldest solution, Prolexic. Still have growth there, still seeing huge attacks just at the routing layer. The biggest pillar is the WAF, which is the WAF platform, which is more than half. And today bot management over $250 million and growing nicely. And then other capabilities built on top of that. And then we talked about the Guardicore Platform, which is the Zero Trust solutions, solutions for inside the enterprise.
Got it. So I want to touch on the delivery business as well, which was kind of the initial core of Akamai. I'm going to open it up for questions after. So if everyone has questions, get them prepped now. So what is the state of the industry right now when it comes to delivery? It looks like overall traffic volumes are maybe a little bit pressured, which we haven't seen historically. But on the other side of the equation, more consolidation of that spend, some of the competitors dropping out of the industry. How are you assessing that overall kind of CDN delivery market right now?
Yeah. The traffic growth right now is in, I would say, a lull. So small, it's growth, but small, and over time we've seen times when it shoots up and times when it's not much growth, and we're in a zone now where there's not much growth. And at the same time, as you noted, just in the last little over a year now, four of our major competitors have exited. And most recently you've got Edgio, which was the combination of Limelight and EdgeCast, two of our larger competitors, went broke in Chapter 11. And I think the challenge in that market has been those kinds of players were selling the delivery at below their cost. And for a long time, Wall Street and investors financed that in the hopes you get growth, you get a market cap, and you get some return.
And I think the realization has set in. Doesn't work if you're selling below cost, that scale is not helping. The good news for Akamai is we have a vastly better cost structure and very synergistic platform with our security and now Compute solutions. So for us, it's a good business driving a lot of cash flow and profitability, but revenues decline. And that's in part because pricing dropped and not enough traffic growth. And so we're also, as we've talked about, being very judicious about how low we'll go on pricing, what business we'll take. I do think the industry has been overdue for a shakeout, that these smaller players that are losing money just trying to show some growth by lower price to go grab some traffic, that that's not going to work now.
And that's good for, I think, the market as a whole to have that shakeout.
Have you started to see any initial benefits from that industry shakeout in terms of pricing? I mean, you guys are proactively self-selecting towards better business. But just in terms of industry pricing dynamics, that had been very onerous forever. Is that starting to ease up at all?
We'll see. I think the Edgio bankruptcy, I think light bulb went off a little bit, that this is you have to be careful who you're doing business with, because the lights get shut out sometimes in a bankruptcy. And that's going to happen there. And we picked up some accounts as a result of that, which is good. And so I think next year the delivery business won't have a year quite like it had this past year for us. And so things should be better than they were this year. And we'll see when traffic growth gets better. There's a lot of video and traffic and gaming still to come online. So in the long term, I am optimistic for that business. Last year was a this year, 2024, a particularly tough year.
The one thing that we have out there that's a little cloudy right now is what happens to a large social media company as a large Akamai customer and whether or not they're banned in the U.S. So that's something that we've disclosed what the impact would be for us if that happens. Otherwise, I think we're going to see a much better situation as we go forward than we saw in 2024.
Got it. So absent kind of what's going on with TikTok, you do think that the delivery business can, with a good traffic environment, can come back to a growth footing?
Oh, I think over the longer term, we're not calling that for next year, especially with uncertainty around a large social media customer. But yeah, I think the dynamics long term are good. And meanwhile, we're generating a lot of cash and profit. And we're using that with our platform to invest in security and now Compute that are strong growth drivers for the company. And also now over two-thirds of our revenue is security and Compute and growing very nicely. So I think we're in a very good position for future growth.
Got it. Understanding. Any questions from the audience before I keep going? All right. I want to talk about the cost structure. You brought it up a little bit. You guys have been very good at being able to continue to wring efficiencies out of the plant, if you will, on a go-forward basis. Linode also had the internal effect of lowering your cost. You're able to sort of bring some of that Compute capability in-house. Where are we in seeing those benefits? Where are we in Akamai accruing that into your operating margins?
Yeah, in part we benefited from moving our own applications out of the hyperscalers onto our platform. That's saved us a lot. We plowed those savings back into developing and enhancing that platform and introducing new security products, in part through acquisition, in part through organic investment to support the acquisitions. I think it's been very helpful for us. Long term, as that enterprise Compute number grows, that's very helpful. It's a tailwind for us in terms of overall operational efficiency and margins.
Got it. So the framework historically had been security is going to be a better margin business than delivery. And I think investors intuitively understand that. There's a lot more value in terms of how you're using that network and a lot more software versus just underlying infrastructure. I think it's a little bit more of an open question in investors' minds on the Compute side of the equation versus what you're doing now with the combined security. And I think in our heads, we think Compute is a lower margin business than overall security. Does that create a sort of structural weight on where the margins can go on a go-forward basis as that Compute business grows?
Yeah. We believe Compute is a tailwind long term to our margin goals. Now, the impact on the P&L is very different. The big cost for security is the developers and the people, the human capital. For Compute, it's the CapEx, which we currently depreciate over six years. And there's some colo depending on if we're in a big centralized data center versus one of our edge locations. But CapEx is a big driver. And you buy the CapEx and then monetize it. So the P&L looks different. And if you're growing, the faster you grow, then the more the CapEx is in a given year compared to where it was the prior year. It's a good news story. But so yeah, I think when you normalize that out, it is a tailwind on operating margins compared to the 30% goal over time.
Got it. Got it. But there's also some near-term investment you guys have made in building out Linode, building out those enterprise capabilities, marketing behind it. There's been a little bit of a headwind to operating margins this year. When do those investments start to yield? Is that as soon as 2025? Or is this a more extended investment phase that investors should be thinking about?
For Compute, it's extended. The number we most care about is 100 million ARR ending this year. We'll probably guide in February for what we want to end next year. But we want it to be significantly larger. And it depends how fast that grows. And once it starts getting critical mass, then it starts influencing the big number, a $4 billion run rate for the company.
Understood. Unfortunately, we're at the end of our allotted time slot, but Tom, thank you so much for joining us and giving us the update on all the innovation and the expansion of the story of Akamai. It's always fascinating.
Great. Thank you.
Awesome.