Okay, good afternoon, everyone, and welcome to the conference. I am Mark Murphy, software analyst with J.P. Morgan, and it is a great pleasure to be here with Dr. Tom Leighton, who is CEO of Akamai. Tom, thank you so much. Really means a lot to us to have you here.
Nice to be here.
Maybe what you could do is, just for the benefit, Akamai has been around a long time, and it's a large company, and I think most of the audience is familiar. But just for the benefit of anyone who might not be familiar or might not be current, could you give us the one or two-minute overview of the company?
Yeah, sure. We operate the world's most distributed cloud platform. We're not the largest cloud company, obviously, with the hyperscalers, but we do have the best content delivery services, the best web security capabilities, and we're growing a very interesting cloud computing capability that is the most distributed and performs better because you can run your compute instances closer to users and the devices and the data at a much lower cost than you would have with the hyperscalers. Today, Akamai's leading business is security. That's where most of our revenue is. Obviously, we run a very large content delivery business. That's where we started the company over 25 years ago, and compute is very fast-growing. Today, two-third of our revenue is security and compute, and the last quarter grew over 20% year-over-year.
So very exciting prospects for the future.
Yeah, those have, those have been resounding successes. Tom, if, if we think back to when Akamai got its start, it clearly was in the delivery business. You're the market leader in that business. As you just described, the business has evolved tremendously, and you now have security, security is the largest component as well as compute. You know, at some level, all those segments are interrelated, and they're... and then they're all connected through your global network. Can you maybe help us to stitch together the pieces so that we can try to understand what the synergies are between those segments?
Yeah. So the delivery and the security businesses are very synergistic. So when a request comes in from one of our customers, say, you wanna check your bank account, you wanna buy something, you wanna watch something or play a game, as that request is coming in, we're making sure it's really you accessing your bank account 'cause maybe it's a thief who stole your credentials. If it's a bot, we wanna know what kind of bot it is, so we can deliver the right response. Then we go and get, or if we don't have, get the content that it is that you wanna see and deliver it to you. And the content will change based on who you are, what your interactions with the site or app have been. If you're a bot, what kind of bot?
The content is actually tailored on the fly to give you the best possible and most secure experience. So that means that our security products, a lot of them are very tightly interwoven with our delivery products, and it creates great synergy, you know, very cost-effective and great performance. It's not like we have to dish off the request somewhere else to apply security, which some companies do, and then come back and then try to deliver it, and before you know it, it's a terrible experience.
So, and I think you touched on this, in the intro, but you have thousands of points of presence, and these are all spread out across the globe in strategic locations. The number of POPs that you have is, we believe, multiples of other competitors, so it's a remarkable scale. Is that a component of this business that structurally is an enduring type of a competitive advantage? And, you know, can you help us understand what it like, what is the kind of differentiated experience that that's gonna enable versus some other provider that doesn't have as many points of presence?
Yeah. So we operate over 4,000 points of presence in 750 cities, 130 countries. We are inside well over 1,000 networks, not peering, inside those networks, which means that pretty much wherever you are on the planet, if you're going to one of our customer sites or apps, you get a very fast connection to Akamai. We direct you to the server that is right near you, and that makes for a very fast experience.
In addition, it's a great first line of defense for security because if you're trying to do something bad, you wouldn't do that, but if there is somebody trying to do something bad, we will intercept that malicious traffic locally, long before it gets to the target or even in the country of the target, never mind the data center where it could swamp the target. So you get great, you know, defense in depth with security, and increasingly, we're doing compute close to the end user. Now, for a long time, we've operated, you know, JavaScript as a service at this edge platform. So in these 4,000 locations, we'll spin up JavaScript apps in five ms in a serverless way based on end-user demand.
Now, what we're in the process of doing is taking our edge, our computing containers, VMs, Kubernetes, and migrating that. We want this year to be in 100 cities by the end of the year. Nobody does that. And over time, that will grow. And by getting compute instances closer to the end user and really at the edge, again, you get a better performance and better scalability. We can also do it at much lower cost.
That's very helpful, and I want to come back to a couple of those aspects that you touched on, including the edge and the Gecko Vision. But before we do that, Tom, I do think we should spend upfront, you know, some material time on the security business. You know, you have these very deep capabilities, and they're embedded in interwoven across the platform. You know, you've had quite an investment posture on the security side. It's been organic. It's been a combination of organic and inorganic, excuse me. And we also know, I mean, I'm sure everyone in the audience knows, security spending is seeing prioritization, right?
That's a function of how intense the threat landscape has been. Thank you for giving our audience the benefit of the doubt that they're not bad actors out there, by the way. But so what is the differentiated role that Akamai brings to the table in security? Can we peel back the layers of the onion a little bit deeper on that, and maybe just help us understand the longer-term vision?
Yeah. So our goal is to protect enterprise applications and data, and we do that today in three tiers or layers of products. The first is protecting the enterprise from denial-of-service attacks, and that typically is at Layer three and four. It's intercepting the traffic that is designed to swamp out a critical resource, like DNS. If you take out the DNS for an enterprise, they're offline, or swamp a particular application or a particular server. And so we intercept that traffic in special scrubbing centers and in our 4,000 POPs long before it gets to the data center, and we have the market-leading capabilities to do that. The next layer of defense is at the border between the enterprise and users.
The users can be consumers, or they could be partners of the business, and this typically works at Layer seven, and so we're trying to stop the injection of malware, the exploitation of vulnerabilities. This is where the bot management capability comes into play. This is where we do special things when somebody logs into a bank account or a commerce account, to make sure it's really you and not somebody that stole your credentials. It's where we stop site scraping, site corruption, theft of data, and again, with the market-leading solutions, you know, it's a marketplace that, in the cloud, we started web app firewall as a cloud service. We started bot management, and today we're the market leader, and this is the majority of our security revenue is in, in this layer.
And then the final layer of defense, which is now the fastest-growing, is inside the enterprise. You know, today the problem is that you could buy everybody's solutions, everybody, and malware is still getting in, and Gen AI is gonna make that worse. The key really is to identify it when it gets in and to proactively block it from spreading. Malware's not so bad when it first gets in. Ransomware, if it gets into a server, who cares? The problem is it then spreads from there laterally to lock everything down. You know, if your data exfiltration malware gets into your HVAC system, that's not a problem yet. It's a problem when it uses that as the base to go collect all your data and exfiltrate it, and then you got a big problem.
There's where you need a segmentation solution to prevent the spread of the malware and to identify and tell you that it's there, and we have the market-leading solution for that with Guardicore, and we've just recently put that on the same agent and same pane of glass control panel as our zero trust network access. So now we combine north, south, and east, west, which is unique in the marketplace.
Yeah, and that's the Akamai Guardicore Platform, right?
Right.
And I wanna come back to, I wanna try to understand that a little better in just a moment. But just to clarify one comment that you made there, because you said generative AI is gonna make the whole problem worse.
Yeah.
I mean, is it implicit in that? I assume you mean that the LLMs are gonna be creating malware-
Yeah
... polymorphic malware and all this.
Yeah.
Do you think that that's gonna happen faster than it's able to... we're able to use the Gen AI on the other side to identify-
It already has.
Okay.
Yeah. No, you can train really nasty malware very quickly, and you don't have to be really skilled to do it. Gen AI can help with QA. We use Gen AI in our new human interface to the Guardicore platform, so you can do human language queries to find out what's going on in your infrastructure, what applications you have there, what an application is, and what it does. Does it have the most recent firewall rules? So it's useful on the defense, but it's much more enabling now for the attackers.
So coming, you mentioned, you mentioned segmentation. You mentioned that your security business is growing faster inside the enterprise than outside, and I think the first part of that you mentioned was segmentation. You have this very strong product called Guardicore, very healthy momentum. You landed one of the top telcos in the U.S., you landed a supermarket chain with more than 1,500 stores, and then a major software company in Latin America. And that, I believe, was all in Q1. What is driving the success of Guardicore?
And when we think about the success there, has that kind of whet your appetite or upped your confidence to say, "You know, this is a good way to go," to maybe buy a security company of that size, that scale, that profile, and then leverage it well?
... if you get the right company and the right product-
Yeah.
We were fortunate, because I think you go back a few years, and I think micro-segmentation was not well understood or appreciated in the marketplace. You know, we felt it was gonna be the important thing to have. And it is working out, you know, in that direction, because it is really the best defense against ransomware and data exfiltration malware, and Guardicore has the best solution. Now, at the time, they weren't recognized as number one, but we liked them the best, and now they are recognized as the best defense.
So, when we think about... Just a very quick question on your go-to-market and, you know, how you're able to drive the rapid absorption that you have seen with Guardicore, and then even on the API side. Our understanding is that there's a bit of a unique dynamic in the security market, where sometimes for this very complex enterprise level of security, the buyers actually wanna take that through, they wanna absorb it through a channel.
Mm-hmm.
And, you know, these companies that'll come up are Optiv and CDW.
Mm-hmm.
You know, you'll get CDW and Carahsoft and Insight and all that. Is this something that's active for you as well, or is Akamai able to say, "We already have the direct relationship with these customers, there's no need for them to go through a channel?
No, channel's increasingly important for us. Guardicore, in particular, is channel only. API Security, we're making investments there. That'll be largely channel. So yes, and security channel partners are very important for us.
So let's talk for a moment, then on the zero trust side. The more companies move into a cloud environment, the more they're gonna be prioritizing that as an architectural advantage for them. One of the questions we commonly hear from investors is that, "Well, you have some pretty big, pretty strong, pretty established players there, right?" And they will commonly mention companies like Zscaler and Palo Alto and others. Can you help us understand, on the zero trust side, what is it that's differentiating Akamai, and why is that a logical adjacency for you?
Yeah, so, well, we have the best solution, you know, for segmentation, best way to stop ransomware and data exfiltration malware, so that's a big advantage for us. We've now combined that with our, the Akamai Guardicore Platform, as we talked about, with zero trust network access. There, we had a very competitive solution, so we would compete with ZPA. And we've now merged them, so you get north, south, east, west on the same pane of glass-
Mm-hmm
... and agent, combined it with MFA, you know, DNS protection, and threat hunting. And that's a unique offer in the marketplace, so I think gives us a very strong, you know, leg up against the competition when it comes to protecting the applications.
Mm-hmm.
We don't, you know, solve every security problem, but if you wanna stop what we think are the big ones, we have the best capabilities.
You, I'm wondering if that helped lead to this win that you had, one that certainly got our attention in the recent quarter, which was with the U.S. Army. You talked about this, zero trust in battlefield networks.
Mm-hmm.
The comment was that it was a competitive evaluation of more than 40 vendors.
Mm-hmm.
So we would have to think, and then the certifications that you would need to and the validation you're gonna get as a result, we would have to think that every security company probably wanted to win that piece of business.
Mm-hmm.
Can you walk us through how exactly, you know, did that transpire? What was it that you enabled to win that business? And I don't know if there's any comment or any way that we can think about the revenue implications of a big win like that.
Yeah, the revenue is important for us, so it's a substantial win, and there's, I think, upside, you know, down the road. We have had very successful businesses with, you know, the Defense Department and the government, in the U.S. and in other countries. I think we have a very strong reputation for very high levels of security and reliability in our products, and I think that was very helpful. Of course, then when it comes to these solutions, you know, they are the best at doing the job.
So then, let's go back, because you mentioned it just a moment ago, the Akamai Guardicore platform is something newer that you're talking about.
Mm-hmm.
I think you mentioned, so there's micro-segmentation, there's network access, there's firewall, there's threat hunting. I think you even mentioned a couple of other elements of that. It's going into a single console. It sounds like it's going into a single pane of glass.
Mm-hmm.
Is that a pretty big engineering effort, and, is that a marker of Akamai reaching a point where you're saying, "We don't just have a couple of, you know, items on the menu-
Right
... we sort of have everything, and so it's, now it's time to put that into one single console?
Yeah, it's not everything, but it's the key things, what we think are the key things. And not just on one console, but one agent, 'cause before they were developed differently as products, and you had different agents, and that's a very expensive real estate. And so to be able to go to the customer and say, "Okay, just one agent covers you, you can now apply your business logic across North-S outh and East-W est," that's very compelling, and customers have wanted that. And it did take time to do. You know, we've been working on that over the last, you know, couple of years, and now live in the market, and very positive reception from customers.
So one agent is gonna be less taxing on all the other resources, that a customer has-
Yeah, 'cause you put it on your application or your devices, and you people don't like sticking stuff there.
Yeah. Okay. In that, can you take a platform that is that large and also be selling that through the channel? Or is this gonna be more of a-
Yeah, that's a channel play.
This is a channel play as well. Okay. Let's switch gears and try to talk about what you're doing in the compute segment. You know, part of which goes back and includes the acquisition of Linode. But that segment crossed a $500 million revenue run rate, I think it was two or three quarters ago, and it looks like it is just imminently gonna hit $600 million, right? So it is scaling well in the run rate, I mean, annualized. And you've been leaning in, you've talked about how the CapEx plan is 16% of revenue on this, on this FY 2024. We've actually seen...
So it was an interesting quarter because the hyperscalers revenue growth picked up for AWS and Azure, and it also picked up for Akamai, where you had 25%, right?
Yeah
... in compute. Are you feeling more convicted now? When I describe that trend line, do you feel more convicted in the compute opportunity? And, you know, should we think of the CapEx as vested investment as... Are the odds going up here that that's gonna see a good return?
Yeah, I think so. I think we were pretty convicted before.
Yeah.
And it's really great to see customer adoption starting now, and ahead of plan. You know, we took up our guidance on compute this year. Last year, did $500 million, raised guidance to be 21%-23% growth, so take us over $600 million. Now, one of the things that is good to understand about that $600 million+ is that there's what we just talked about at the end of Q1 being a $50 million ARR, which is enterprise applications under contract on our compute platform. The other $550 million is legacy Linode, which is small and medium business, legacy Akamai, compute kinds of applications.
It's that $50 million, which is why we bought Linode, why we've made such massive investment around it, where we're going up to go compete with the hyperscalers to take a share of that market.
Mm-hmm.
We want that $50 million to grow to billions. That's where the CapEx investment is going. We announced a lot of, you know, customer wins on the call and the kinds of things that we're doing. That $50 million, at the end of the quarter, was up 4x year-over-year. We're starting to see real traction now, which is great and a little bit ahead of schedule.
Should we think of that, Tom? I could probably use a little bit of help personally understanding what exactly the difference is with those applications. But the way, in my mind, the way we've been thinking about it is, you're taking this compute strategy, you're focusing on swim lanes and verticals that you know very well, right?
Yeah.
As an easy way to get it started, media, e-commerce, retail, travel and leisure. You've got, we know you've got big financial services customers, and so you've got these companies that you've known for a long time. Is that, is that something, is that greasing the skids on these, on these discussions? Is that kind of accelerating this path where you have this, where you have the $50 million?
Yeah, very helpful, because in those leading enterprises, we know the CXOs very well. They know us, they like us, they've used us for a long time. They've wanted us to get into compute. We save them money, and we give them better performance. Now, it's not easy to flip the switch and do the lift and shift, that takes effort, but it's well, well worth it. And now that we've gotten the Linode platform up to a grade that major enterprises can safely use it, we're in a position to start selling it. And there was a ton of work that we've done, you know, in building out the capacity, the reliability, the security, getting all the compliance work done, getting the ecosystem, you know, sufficiently so that you can actually move your applications. In fact, Akamai is our biggest customer.
We don't charge ourselves, you know, revenue for it or account revenue for it—but we're the equivalent of a $100 million-a-year customer. You know, our bot management, account protection, a lot of the apps that, you know, we've been using the cloud before are now on our cloud. And we've seen improved performance and saving a ton of money.
So, can you help us understand what the pricing differential is? Because you said you're helping customers save money. We've certainly heard anecdotes where the Akamai pricing on compute could be 50%-75% cheaper than AWS. I have no idea exactly how to validate that or to understand how common it is or what the ranges are. Could you help us to understand that? And I think, Tom, the other question behind it, it probably a lot of people in the audience will say, "Well, how is that possible?" How is it possible, margin-wise, to come in at such a large discount?
Yeah, so that would be in the case of a, say, a big media company or maybe a commerce company that's very chatty applications, 'cause you pay a ton to the hyperscalers, you know, and for those kind of use cases, and we're very good and very efficient at moving data around. Obviously, we're the world's largest CDN. Two of the hyperscalers use us to do that. And so that puts us in a very good position cost-wise. Plus, the hyperscalers are charging more in their unit pricing. Prices have gone up, and that has created a lot of room, you know, for us to come in and say half the price for a media company and do it at very good margin for us. So that's how we can do it.
So partly, it's by choice, you know, we're choosing to come in at a lower price point, and partly, we have a really good cost structure ourselves for those kinds of applications. Now, applications that you're not moving data around or pretty static, yeah, probably not 50% in that case. It'd be cheaper, but not, you know, the big discounts.
And what is the scale of the spend that you can target on the, on the compute side of the ledger? Because again, and I'm not an expert on this at all, but we've—you know, our understanding is there can be cases where if you look at, say, especially for a streaming or a media company, and you say, "Well, what do you spend on CDN?" That's a little bit like looking at the tip of the iceberg. It's the part of the iceberg-
Yeah
... that's above the water, right?
Yeah.
Then you say, "Well, what are you spending on the compute or the compute/storage?" Well, that's the rest of the iceberg that...
Yeah
... is under the water, and it and it's just a lot larger. Can you help us? I and people will say it depends on the encoding and transcoding and, like, exactly what's going on in these use cases. So is it a way to conceptualize what's happening here as you go from CDN to the compute side with these large customers? You're going from tip of the iceberg to the part that's underwater.
Yeah. So for a big media company, they'll spend 10x on compute and storage than what they spend on delivery and security. And it'll be, for us, at higher margin. So yeah, huge potential market for us.
You're equally equipped to take on the compute and the storage?
Yep.
Okay. So, what else did I want to try to get into here? Okay, so if, if we go back to the discussion around what happened with AWS and Azure, you know, you'll see, well, the growth is picking up, but it is still happening underneath a mindset where they're very cost conscious, and they're gonna be very deliberate, and they're not gonna be moving all that rapidly to spend money on anything. Kind of a new normal around, you know, just where we are with interest rates and cost of capital. Is that something... So first of all, do you see that? Do you see that type of behavior?
Secondly, is it something, is this something that can work in your favor?
Yeah, it depends on the vertical, but obviously, media companies are pretty stressed, financially-
Yeah
... want to cut cost. Commerce companies, many of them are. And the fact that they need to cut cost, and they'd just as soon stop spending a lot of money with their biggest competitor, does help us. And the fact that we can help them cut costs and actually give them better performance, that's a big plus. Now, it's not easy to lift and shift. It takes time. It took us a lot of time to move our applications out of third-party cloud, you know, into our, onto our platform. But, yeah, the environment is helpful for compute. You know, the same environment that's causing big media to want to spend less on delivery, helps us when it comes to, you know, selling compute.
So we're down to about 5.5 minutes. I just want to do a quick check with the audience here. If you have a question, go ahead and raise your hand, and we will then run a microphone to you. Anything out there? No. Okay. So you did mention this concept as well of eating your own cooking or being your largest customer, your internal, your largest user. I thought you had said the goal was generating $100 million in OpEx savings coming from the external cloud providers. Can you talk to us about how... I wasn't sure, based on how you were describing that, how far down that road you have gone. Are we halfway complete? Are we closer to the end of that already?
Has it gone pretty quickly? And are there any learnings from it? Can you take what you've done and say, "Okay, this is what it's like to move a workload over," and then you can apply those learnings to your future customers?
Yeah, it is very helpful to do it yourself first.
Yeah
... 'cause you know what your customers are gonna go through. We also did a lot of development along the way that now our customers can benefit from. And we're more than halfway through at this point, and by the end of this year, we'll be mostly done with it, and very large savings and improvements in performance, and development of capabilities that our customers are gonna benefit from.
So, Tom, let's spend a couple of moments talking about the delivery business. You know, it's been important historically. You've been very thoughtful, you've been very deliberate because it is a business that has very high margins at Akamai. It's a very profitable business. And then, you know, but the focus has been shifting more to security and compute, and, you know, I think the sense that we have had is that you wanna try to maintain the strategic portions of the delivery business.
Mm-hmm
... and, maybe let go of some of the other pieces, depending on the economics. Can you give us a sense of what the long-term state of that business is gonna look like? I mean, is it a perpetual decline in terms of the mix because of the way security and compute might grow, or is there a point where it would level out?
Well, as revenue from the delivery business, I don't know why it would be a long-term decliner. You know, it should stabilize. You know, the market pricing, you know, continues to have a lot of irrationality to it. You know, and this year on top of that, you know, a very large, one of our very large customers has had their own challenges, geopolitical challenges, you know, that they're dealing with, and that's had some adverse impact to us. I do think it becomes a smaller share of our total revenue. You know, you've got compute and security at two-thirds of the revenue growing rapidly, and this year, delivery declining. I think longer term, it should stabilize in revenue. Shouldn't go away, it's a vital capability that major enterprises need, but it does become a smaller piece of the pie.
The good news there is that when you do have these occasional dips in the traffic, it becomes much less noticeable and doesn't, you know, impact people's view-
Yeah
... of the company.
Okay, and then and so there was a dip in the traffic. Then, I think on the earnings call, you mentioned lower than expected traffic in delivery in over a two-month period, and it seemed to be most notable in gaming and video.
Yeah.
There was a comment about the drop in load and active users on some of the streaming services in April. What do you think is potentially, you know, driving it? I know, because we know you had, or Ed had referenced, you know, crackdowns on password sharing, and then I think some of the ad-supported, you know, initiatives that are out there. I think we're trying to understand. We're trying to understand if that could, we could look back and say that was a bit temporary, or is that in some sense macro catching up with, you know, some of your customers, that they're taking those steps?
As best we can tell, yeah, and it's anecdotal. I don't think there's any big core driver, and it's normal. You know, in the delivery business, traffic generally rises, sometimes rises faster, sometimes rises slower, slower. There can be dips. You know, the biggest impact to us was one very large customer who's had some adverse things happen to their business. And so they're in a cost-cutting mode, and we've also talked-
Are you talking about the social media customer?
Yeah.
Yeah. Okay.
That's the biggest-
Yep
... issue for us.
Okay.
That was the majority of it. You know, there's anecdotal stuff where traffic, traffic growth a little bit less, and that's normal to have that kind of thing happen.
So Tom, in the final moment that we have here, what are you most excited about as we progress through the course of this year, and then you start thinking forward to the next year?
Yeah, I think security and compute. Security, you know, Guardicore becomes just even more core and important, growing like mad now, but, you know, very successful. We're really excited about API security. Every CSO I talk to knows they need to have some idea of what APIs they have exposed and some security to protect them. They're vulnerable right now. And then you look at compute, enormous revenue and profit opportunity for us. Just, just enormous. Now, it takes us time to get it, but we're off to a good start this year and ahead of where we thought we'd be. We got the whole field now out there selling it with specialized support, so just a huge upside for us.
Okay, great note to end on. Tom, I can't thank you enough for taking the time to be with us here.