Akamai Technologies, Inc. (AKAM) Earnings Call Transcript & Summary

Excellent. So let's get kicked off here. Welcome to the Morgan Stanley TMT Conference. My name is Keith Weiss. I run the U.S. software research effort here. And very pleased to have with us my first presentation at the TMT Conference, Tom Leighton, CEO and Co-Founder of Akamai. So Tom, thank you for joining us. Before we get started, a brief disclosure. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

All right. So with that out of the way, let's kick into what has been a really exciting year for Akamai. It's always an exciting year. But I think this year, in particular, that we're talking a lot more about a connected cloud service. So it's something that you guys have really brought into the forefront of the Akamai story over the past year. I think it was last month, you unveiled the connected cloud service, integrating Linode, which you acquired into the broader Akamai 4000-plus edge computing locations. Can you talk to us about why this is so exciting from your perspective? I mean, Akamai has always been kind of your baby and sort of what you've been growing. I can tell you're very excited about this opportunity, probably more so than other ones that we've seen in the past. What is it that you see here? What's the unique opportunity for Akamai to really, in a differentiated way, address this market?

It is exciting. It's a chance to provide a complete cloud service to our customers. And my excitement is derived from our customers' excitement. Think about it is you've got 4,200 edge locations. These locations deliver content, they do edge computing, JavaScript, Function as a Service right near the end user in the last mile. They also do the first layer of security. So they intercept the attack traffic before it gets anywhere near the data center. And then there's 2 stages inside of that. At the very core, we now have core cloud compute and storage. So that you can run VMs, containers, monolithic, object storage, block storage, everything you could do in a hyperscaler. And we're building out an intermediate layer, we call distributed computing. So I think containers as a service, where today, you can't get that from a hyperscaler because it will be initially in multiple dozen locations ultimately and maybe a couple of hundred locations, all connected with the Akamai backbone. And a lot of people don't think about Akamai as a backbone, but we run one of the largest backbones in the world, which can also integrate directly to our large customers. And so now they'll be able to do everything on Akamai. They will be able to build their apps on Akamai, have their storage on Akamai, run the apps, distribute them, get great performance and, of course, have security, have it all protected by Akamai. And having it all in one platform makes it a lot more efficient. The users get a better experience and the costs are a lot less. And this is a time when cost actually, I think, matters more than usual. And so our customers are very excited about this opportunity.

Got it. So it makes a lot of sense, sort of that one platform perspective. And you're right, everybody does push for kind of lower cost options, particularly with some parts of the equation like core infrastructure, like compute and storage and like. But are there specific use cases and workloads that you think Akamai is uniquely suited to address with the connected cloud service? And maybe what are they?

Yes. workloads where scale matters, locality matters, performance matters. Think about this in some of our large verticals like media, gaming, commerce, those companies' performance matters a ton. And with this capability, we'll be able to give the utmost in performance. In fact, our first early adopters of the cloud service are big media companies where you need to do the transcoding and special stuff for an individual as they're watching a video to make it really high performance, you want to do that close to the end user. It's a lot more efficient that way and you get better performance, could be transcoding, could be uploading. One example is you just created a video, you want to upload it. So all your friends can see it and do that in real time. And that's something you want to do close to the edge, either in containers in the distributed edge or actually right on the edge server. Commerce, same thing. Think about a commerce site, you've got a massive amount of data that stored in the core, you're doing data processing on that. But then a user comes to interact with your site. And what you want to do is you want to take the data specific to that user and get it out closer to the end user. So to make a decision, you don't have to come all the way back to the core and enter this monolithic database, you can do that now in a container much closer to the end user. And then, of course, on the edge, you want to be doing quick processing of what that user is doing right now. And then maybe some of that comes back to the middle region to do some compute based on the past history of that user to make sure you tailor the content and get a great experience for the end user. And having this all together in one platform just makes it a whole lot easier, faster. And of course, when it's fast, you buy more. And so our customers are happier.

Got it. Going to go a little bit off script here. But when I hear that kind of dynamic, it's really hard to go through this presentations without sort of trying to understand sort of the AI context, kind of what we're speaking about. And you talked about a little bit on the last conference call about this platform could be good for supporting those AI workloads. And when I hear you talking about sort of what would work well on the edge, it seems like, yes, maybe you're going to train the model in the centralized location, but you're going to want to run those models close to the edge and do that inference of where you could have the fastest response time.

That's exactly right. So you're big monolithic processing, that's asynchronous, or offline, done in the core. That's great, massive data there. But when the user is interacting with the site, you want that local and on the fly and really quick. And more and more of the APIs today are very chatty. And so the round trip time matters because you're doing a lot of that to present something to the user and conduct a transaction.

Got it. So when you start to talk about the use cases, you started with media and gaming. So I'm assuming that you're starting -- you're having these conversations with your core customers and customers that have been with Akamai for years, if not decades. Can you talk a little bit about their initial response? How interested are they in augmenting what I'm sure they're already doing in other cloud services with what you're bringing to the table with the connected cloud?

Yes, there's a lot of excitement. Those companies already trust us to be reliable, to scale for security. They trust us with their secret keys, and they've known us for a long time. And now we're in a position to offer the last piece, which is the core cloud compute and put it all together in a way that makes a lot of sense and also to save money. You think about it today, we're the front end for those sites, by and large, all the data is passing through us. And then today, we end up passing it off to either their own data centers or to a hyperscaler, which involves a lot of cost and some inefficiency. And then, of course, when you want to get something back out of a hyperscaler, that's even more cost. And now that goes away. So it makes it not only faster and more efficient, but really much less expensive. And in media, in particular, we're seeing the big customers, they're in a lot of pain right now in terms of their cloud bills. They're actually rising quite a bit and taking a bigger bite out of the budget. And we can help with that problem. And so there's just a lot of interest. In fact, I think one of the firms did a study of 50 of our customers saying, "Hey, would you buy this service, Akamai Connected Cloud? And I think a vast majority said, yes. So we're pretty excited about the prospects.

Got it. And then can you talk to us a little bit about the pricing strategy because lower cost is definitely a part of the equation. I think you guys just recently made announcements about egress fees with Linode. So I guess the overall question is how are you guys thinking about pricing strategy? Is it specifically meant to be -- this is a lower cost alternative to the hyperscalers?

Yes. I think generally, it will be a lower cost. You just look at our list pricing and it's substantially less than what the hyperscalers would charge. And of course, if you are a big egress kind of entity like commerce or media or gaming companies would be, there's even more savings to be had. So it's substantial. And our typical customer in those verticals today will spend 10x in cloud cost and they'll spend with us. So it makes a big difference when we can include that as part of our capabilities.

Got it. Do you see a response from the hyperscalers because -- when you talk to whether it's AWS or Azure or GCP, they all talk about the edge as well, right? And they talk about it in various forms, whether it's AWS talking about telco partnerships and Outposts or Azure just being a more distributed network it themselves or talking about the on-premise environment being kind of part of the edge. What do you think the competitive response, not just to Linode and Akamai in this marketplace. But there's other vendors, whether it be Cloudflare or whether it be DigitalOcean, who are trying to sort of push on this edge, do you see any like real competitive response from the hyperscalers?

In the last few years, everybody has been talking about edge as a marketing term. I think only Akamai is really at the edge. I mean nobody has thousands of locations that do function as a service, that do delivery and that do security. We're really unique there. And it's not easy to create that. So everybody talks about it, but nobody really does it. Now in terms of the hyperscalers and competing there, a couple of the hyperscalers are among our biggest customers because we offer the best when it comes to delivery and performance and security, even though we compete with their own services there. They're big Akamai customers. Also, Akamai has a little bit of an advantage in that I think we are much smaller as a company than they are. And so it's like we'd like to get a few percent market share in cloud compute, which is enormous for us. That's billions of dollars that we'd add. But it doesn't matter to them because they're growing at a very rapid rate. And if we're successful in over a period of a few years, you can get 1%, 2% of market share that they don't even notice that, I would say.

Got it. And then on the flip side of the equation, I won't press too hard on the margin side equation because you're CEO, not the CFO here.

No, no, that's all right. Margins are really important.

But when building out networks overall, as you know, is expensive and building on kind of a new network with the cloud network is going to take more investment on the CapEx side than the OpEx side. Can you talk about the investments upfront necessary to kind of build out this network? And how investors should think about the CapEx intensity of the business on a going-forward basis? Does this fundamentally change it from what we were talking about 2, 3 years ago when it was more so sort of a CDN business and a security business?

Yes. So it is more CapEx-intensive. And this year, you really notice it because we're doing the initial build-out. And so there are several hundred million dollars there. Now a big chunk of that is for our own use of the cloud, which is going to save us a ton in OpEx going forward. We're going to spend ballpark $100 million this year. Already, that's less than we would have spent without our own cloud capability, but the vast majority of that, we want to have go away next year. But so this year, you see us spending the initial build-out in CapEx, and we don't really get the benefit of that in revenue or OpEx savings starting -- until really starting next year. And so that's very noticeable to investors this year. But after that, it's a much better situation because you don't spend a lot more CapEx until you fill the CapEx you had. And by then, you're generating a lot of revenue at very good margin. And now it's all about growing the business.

Got it. Got it. And then just one last question on Linode and you guys have. There's a couple of other parts of the business we want to get to as well. Can you talk about that growth expectation of what have you guys kind of templated as kind of initial growth out of the gate for Connected Cloud? And where do you think it could go to over time?

Yes. So Linode, of course, was targeting small and medium business and the developers. And we're maintaining that business growing, say, mid-teens, but what we're really focusing on is adding large enterprises. And that means upgrading Linode's capabilities, both in terms of what they can do and also the scale and distribution. And so that's what we worked on last year and are working on this year so that we can go after the world's biggest enterprises for their mission-critical applications. And the goal is by the end of this year that we can be signing large deals with major enterprises, and that would generate revenue going to next year. Now in terms of overall, the business, last year, we did $400 million in compute roughly. This year, roughly $500 million. But the growth is really going to be driven by business we don't have a lot of yet, which is big enterprise, mission-critical applications, and that's where we're going after ultimately billions of dollars of revenue. So that's something we want to be growing very rapidly, really starting, I think, with revenue next year, and that drives the whole compute business going forward.

Got it. Got it. That makes ton of sense. So I want to shift gears to the security business. Security is now a $1.6 billion run rate business for Akamai. It accounts for about 40% of your overall revenues. And you guys have done a remarkable job of building this business both organically and inorganically, leveraging sort of assets you have out there to both get into customers, but also sort of utilize the underlying network that you have in place. From an investor perspective, while there's a ton of excitement on the cloud side of the equation, does that necessarily mean that the security side of the equation has to take something of a backseat that this is now kind of like the secondary growth driver for Akamai and not as sharp of a point of focus in terms of where you're going to be driving growth from?

No, we've got a ton of focus on security. In fact, this year, pretty remarkable in a sense, security will be our biggest source of revenue, bigger than delivery and CDN, which is a remarkable shift for a company that started the CDN space and has a major presence there. So security incredibly important and very strong growth driver for us. And I think it works well in conjunction with the compute business and the delivery business. Now in terms of the growth drivers there, we've got the biggest -- the web app firewall. And that market, we penetrated it pretty well. There's room for growth there, but what we're really seeing the growth is the capabilities on top, things like bot management. And today, for a lot of our customers, most of their interactions are with bots. Most of those interactions aren't a good thing. The extreme example would be like a Taylor Swift concert, which is an example of what can happen if you're not using Akamai Bot Management. On top of that, account protector. And that's all about identifying that somebody is trying to access an account, a bank account, commerce account, media account even. Where they have the right credentials, the log in and the password but they're not the right entity. Maybe it was stolen or something else. And we're really good at identifying that. And that's a product that now is getting very strong ARPUs and adoption, which is great to see, also things like Page Integrity Management, which is all about identifying and blocking the bad effects from malware in the digital supply chain, things that impacted, for example, British Airways, Magecart like attacks. And now it's a requirement in the PCI standard. If you want to be PCI compliant starting in early '25, you've got to have something like that, and we're in a really good position to help customers there. Most recently, brand protector, all the phishing sites that are out there and commerce sites, bank sites are pretty helpless trying to find them all. We had a recent test with a large bank in Japan. And in 1 week, we found 250 phishing sites, not 1 or 2, 250. They knew they had a problem. They had no idea how bad it really was. So there's a lot of capabilities we're building on top of the web app firewall. API protection is another one. A lot of the transactions today are API-based. You need to have special protections there. So I think a lot of room for growth with services that are relatively new that we're bringing to market, areas that we're creating, the first to provide real defenses for growing rapidly on a smaller number. So it takes a little while for them to raise the whole.

I think you did give an indication on that, Bot Manager being one of the faster-growing sort of modules you guys have heard. Any sense you could give us of the scale if we think about the sort of on top of -- solutions on top of the WAF like? What kind of scale are we talking about today? And where could that go over time?

Well, so one of those, as it's been out a few years, a couple of hundred million dollars growing from there. They won't all get that big, some will get maybe bigger, but it takes time to do that and get the adoption. And also, cross-selling with our enterprise security suite, a lot of the Guardicore customers don't yet have our application protection services and vice versa. Most of our customers still have one security product. So we're working on cross-sell. Also, security channel is a very important area, especially in enterprise security.

Got it. And can we switch gears and talk a little bit about that. The enterprise security side of the equation because that's the newer part of kind of the security portfolio. It's something you guys have been investing a lot in both on the product side of the equation as well as the go-to-market. Maybe you could touch both sides of that. Like what are the sort of the state of play on the product side of the equation? Where have you made the recent investments? And where are we in building out that more enterprise-focused sales force that could go in and do those enterprise security sales?

Yes. So the lead product is Guardicore, which is the micro segmentation. And I think in a lot of ways, that's the most important defense and enterprise you can have. Today, you can buy everybody's products. You can buy ours, you can buy all our competitors. And the problem is malware is still getting in. It's just really hard to keep all the doors and windows really locked, just malware gets in. And the key really is to quickly identify it and block it from spreading. And that's what micro segmentation does. And Guardicore is now recognized by the analysts as the leading solution for that. And it stops the damage from malware, stops the damage from data exfiltration attacks. Basically, it puts a software agent, a little mini firewall on every application. So you can see what's going on. And if an application starts doing something unusual, now you're going to alert it, you go investigate it. And you proactively block the communications. You don't want your HVAC unit communicating with some data repository as an example. And so that keeps the malware from spreading, the ransomware from getting everywhere, so it can lock everything down, also gives you great visibility. If a zero-day comes out like Log4j, we can tell our customers within hours everywhere they have the vulnerability. And without that, typically, it took weeks or months to go find it. So I think Guardicore is the flagship product. And then around that, we bring in other capabilities like Enterprise Application Access, which is often called the North South to Guardicore's East West or app-to-app being what Guardicore does human device employee device to app being what application access does.

Got it. Got it. So Guardicore has made a lot of good progress thus far. On the enterprise access side of the equation, that's a -- I mean, they're both very competitive spaces. But that's a competitive space we hear a lot about, a lot of different vendors going after that sort of access side of the solution. Can you talk to us about who you're competing with there? And sort of what's the Akamai advantage when it comes to the access side of the equation?

Yes, there's -- with Guardicore, there's the classic approach to segmentation, which was physical. So you would physically partition your network. And you'd put firewalls in between the partition, which was very inflexible. And so a lot of companies didn't do it. Those that had to have it by regulation made very big partitions because it was so inflexible. And of course, if you have a big segment, you're going to have big damage when the malware gets in because it can go across a large portion of the enterprise. So that's not a very good approach. Using the agent-based approach in software like Guardicore does much more flexible, easier to implement. And there, I'd say the leading competitor is Illumio. In fact, in the last year, we've overtaken Illumio in terms of being the recognized best-in-class capability.

Got it. And then just on the distribution side of the equation. And I want to make this a little bit broader than just security but include security. If we think about -- whenever I think about distribution from a specific company, I'm always thinking about how do they go across an organization, how do they effectively go across. The bread and butter of Akamai historically had been sort of the website operators. And the people are ensuring the performance of the security of those websites. Security and particularly enterprise security seems to go further afield from that, especially when you get into access. How do you make that leap? And then on the cloud side of the equation, you need to talk to developers about where they want to deploy their software and the administrators who are going to keep that up and going. So can you talk to us about the go-to-market strategy? And how do we make these hops from your traditional customer bases to the new buyers for the new solutions?

I think the biggest leap was when we went from delivery to web security. That was a big hop, which we've successfully done. The hop from web security to enterprise security is smaller because we're already dealing with the CISO and the Infosec organization, different buyer at the low level, but roughly closer than going from delivery to security. Now for cloud, delivery to cloud, smaller hop there. The entity that I think big media and big commerce that's putting all their assets on us for delivery, we're dealing at a pretty high level there on the CIO generally organization. And the compute side is not far away. And we're a big enough spend with those folks that we're dealing with procurement. And suddenly now we're able to talk about helping them with 90% of their spend or 95% compared to what they're spending on us, and that's a conversation that now is very timely.

Got it. And then where do partners come into the equation? Because historically, when it was just a delivery business, it was very much a direct focused model. And even with web security. But now it seems like you're looking to engage partners more when it comes to enterprise security, engage partners more on the compute side of the equation. Can you talk to us about sort of the evolution of the go-to-market to now and food partners and where we are in that evolution?

Yes, almost half our revenue is partner-driven today. On the enterprise security side, it's our partner. Guardicore is 100% partner. You're right that with core basic delivery less so. I think with compute, there's a lot more room for partners to add value. The effort involved in a lift and shift, for example, either out of a data center or out of a hyperscaler that takes some effort. Also managed services going forward. So I think that's going to be -- partners will be very important on the compute side.

Got it. Got it. I do want to open up for questions, just see. I'll ask one more question while you get ready. So if anyone has questions, feel free to raise a hand. I'm going to sneak in one last one. And it's just about M&A and the Akamai has used M&A to sort of expand out solution portfolios. Obviously, Linode and Guardicore were 2 of the larger acquisitions. How should we think about the potential for further M&A on a go-forward basis? And how do you guys think about what's a good target for Akamai, either from a sort of market perspective or sort of like the company that you're going to acquire?

Well, I think Guardicore and Linode were transformational. They were larger scale. And it's unusual, we do one of those lot alone, too. And so I don't see us doing something like that again in the near future. If we found the right opportunity, maybe, but I think less likely, more you'll see tech tuck-ins, maybe from time to time, something that's more of a roll-up nature of the economics are really attractive. But generally, I think technology capabilities that we can incorporate into our platform, probably focused in security, maybe a little bit in compute, but smaller scale acquisitions over the next couple of years.

Got it. That's super helpful. I think we had a question up front.

Can you talk a little bit more about the investment for AI at the edge. You talked a little bit about making the CapEx investment that you'd be well positioned for it. But just given what we've been hearing a lot about the kind of hardware required to run inference for some of these large language models. They require A100. Those are backordered, H-100 might be good for it. Is that something that Akamai is looking into? Or can you talk a little bit about that?

Yes, we used it, of course, in our own services or potentially around security, but by and large, it's a driver of the need for compute. And so that's something that drives the whole industry for compute, which we would also benefit from because it's just very compute-intensive. We don't have our own special services around AI, if that's what you're asking.

Any additional questions? No. Thank you, Tom. Appreciate the time, I just want to connect maybe more so thinking about -- I was thinking about CapEx intensity and our success in the field, how should we be thinking about if we overachieve on compute, how that dials down to CapEx and also kind of just the relationship there?

Well, if we overachieve on compute revenue, that means we've filled up rapidly the servers we're buying now because we're getting the initial footprint in place. And that means we'll then spend more on compute servers for, say, next year. So that's the good news scenario. The bad news scenario is it takes us longer to fill those servers and then our CapEx costs go way down for next year. So it's really a demand-driven situation. For $1 of CapEx spend today, very high level, that's $1 of revenue per year once we fill it up. And so the faster we fill it up, the more we would be buying more servers to grow the business, which is the scenario we want to be in. Then you'll see faster revenue growth and better margins in the whole 9 yards.

That's great. And I guess data compute does seem very compelling a lot from a lot of the stuff that we've discussed here. So as we think about those large customers or even those large media customers that want to move over, what, I guess, are they either looking from us in terms of a product perspective like milestones they want to see before they invest more meaningfully put more compute onto the platform. What should -- how should we think about what they're trying to see first before they go into that?

Great question. Build-out is number 1, 2 and 3. And that is more capacity and scale, which we're doing in more locations to make it distributed, so it's closer to their consumer or their end user. There are also some capabilities that we're integrating things like VPC, availability zones, we have database as a managed service, but -- some capabilities there, but we're on a good path to getting a lot of that done this year. So that we'll be able to start really taking on that business towards the end of this year.

Got it. Unfortunately, that takes us to the end of our allotted time, but Tom, thank you so much for joining us. Always an interesting conversation and very exciting stuff going on at Akamai right now.

Very good. Thank you.