All right, I'm Sanjit Singh. I am doing my last presentation for TMT 2026 , we're gonna end it strong with Akamai. We have Ed McGowan, Chief Financial Officer, at Akamai. Ed, thank you for joining us at TMT 2026.
Well, thanks for having us. Great to be here.
There's a lot to talk about. The Akamai story's gotten really exciting, and so we're gonna dive into all the things that's going on in the company. Before we get there, for important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosure. I wanna start the conversation with you, Ed, on the various acts that the company-
Mm-hmm
... has been going through, right? The company sort of pioneered the delivery networking market. You're sort of the market share leader there. You entered into an act two, which saw you build a $2+ billion security business over the last decade. Now in recent years, you're pursuing act three, which has seen the company enter the public cloud market and more recently, the GPU as a service for edge AI inferencing with Akamai Inference Cloud. A lot going on. For those investors who haven't been close to the story for some time, can you lay out the vision here for Akamai's act three and why the management team is excited about the prospects, prospect for growth heading into the AI era?
Yeah, great question. You know, most folks know us as for our first act of, as you call it, with delivery and, you know, we've always had a vision of some, You know, we started off actually edge computing way before it was even a thing, right?
Mm-hmm.
This was back I can remember when I first started in 2000, we were doing, you know, some form of edge compute. It was pretty limited in terms of what you could do with it. It was more functions as a service.
Mm-hmm.
The notion of leveraging our platform for compute was always something that we envisioned doing. Security was a very natural extension, so as we're delivering these websites, whether it's a commerce site or a banking site, there's so much malicious activity that happens. It was very natural to be able to protect those sites and do it in a way by leveraging our platform, and at the same time, I'm delivering you a shopping cart experience. I'm also using a web application firewall rules to protect that particular interaction. That business has been phenomenal for us and, you know, grown. Been adding $200 million-$250 million of new security revenue pretty much for the last decade, really. It's a lot of internal innovation.
M&A has been part of our strategy. We've been very successful integrating companies into our portfolio and getting a larger and larger share of the security wallet, and we continue to expect to do that. We actually think that where the next generation of the web, if you will, or the agentic web, a lot of our security products will be applicable there.
Mm-hmm.
Obviously, we'll have to do some innovation and, you know, do some things to conform with the new protocols and things like that. Noname and API Security is a perfect example.
Right
... of that, where you've got massive amounts of data flows. Today, it's APIs. Tomorrow, it will be talking to different models, going back and forth to different systems, and all that will be exposed. You know, I think we have a great opportunity there to continue to grow our security business along with our compute business. There's a natural adjacency with delivery, compute, and security.
Mm-hmm.
Most recently, over the last several years, what was happening is Some of our bigger customers, more innovative customers, are coming to us saying, "We wanna run this application, and it requires us to have root access. We can't just do functions as a service. I can't use your kind of EdgeWorkers platform to program in JavaScript or program in WebAssembly. I actually wanna run my own proprietary code." I'll give you an example. We have a customer who's using us with their own proprietary technology for live streaming, and they wanna do things like synchronization to make sure that the video is synchronized across all different platforms.
Mm-hmm.
They have some code that they use to get much better performance out of that particular stream. In order to do that, they needed to be able to run their code on our platform, and we're not designed to really do that. We bought Linode several years ago, and we, you know, deployed our Managed Container Service, where we're essentially slicing off a sliver of CPU from all of our servers around the world, and we have containers where they can actually run that code in as many locations as they want to. We also saw at the same time as customers were coming to us for these more traditional compute use cases where the public clouds weren't giving them the type of performance that they needed.
you know, one good example, we have a case study on our site about Apple, what they do with Private Relay.
Mm-hmm.
Where, you know, you couldn't have, say, four or five big clouds backhauling all that information and actually run the service because users are everywhere. At the same time, with our security business, we were spending a fortune on third-party cloud, right? If you think of the massive amounts of data that we're processing. We realized that there's a pretty good opportunity here for us to leverage what we do really well, running a massive distributed system, and get into the compute business.
Mm-hmm.
Number one, it's gonna save us a fortune. We've started off as our first major customer, and we were able to offload a tremendous amount of workloads off of the public clouds. That's continuing to grow today. We really proved out the fact that we could run this. We needed to acquire some technology to help us get there. We bought Linode, which was primarily focused on small, medium business, the hobbyists, if you will, and we turned it into an enterprise-grade system. Just so happens now with AI, the demand for compute being in a distributed fashion is even growing greater and greater and did some interesting work with NVIDIA back in the summer, early fall, and just most recently deployed GPUs in what we call our Akamai Inference Cloud.
Mm-hmm
... is sort of the latest innovation. It's a very natural extension of what we're doing, well informed by customer demand, and it's been growing phenomenally well for us.
Like from my perspective, I see stability in your delivery networking business, in your security business, while you're seeing accelerating.
Mm-hmm
... growth in your compute business and, the CIS business. Let's dive into the compute and the cloud infrastructure services opportunity, starting with public cloud. Public cloud is growth is accelerating across the industry, including for Akamai in your Linode business. What are the reasons why this part of the business has moved to a higher gear recently? Is this growth being fueled by a handful of customers, or are you, are you happy with, the breadth of?
Yeah
of that contribution?
Yeah. This actually came up in my last meeting with a group of investors where we were talking about if you had asked me three or four years ago when we started this, we're now at a $400 million run rate in the CIS business, our cloud infrastructure service business. I would have thought we would have a much higher customer concentration. We would have gotten a few big customers to spend tens of millions with us, and then we would just start to get into a much deeper breadth across, you know, say, finance and retail and.
Mm-hmm
... travel and all that sort of stuff. It's exactly the opposite, which is great. We actually, if I look at the composition of that's, you know, hundreds and hundreds of customers. It's anything from a customer doing, you know, I don't know, $200,000 with us upwards to $2 million a month, but it's not driven by a small handful of customers. As a matter of fact, very little revenue today is coming from these mega deals, right. These are starting to ramp up, and they're becoming, gonna become a lot more material to revenue over the next, you know, year or so.
Mm-hmm.
We're also seeing our pipeline full of anything from customers working with our partners. We have a, like a, kind of like a store, if you will.
Mm-hmm
... where you can go and say, if you're a media company and you wanna do media workflow, we have media workflow partners where you can leverage their technology, use our storage and our compute. That's been going phenomenally well. Observability is one of the fastest-growing sort of use cases that we're seeing where customers are using an observability partner of ours and then ingesting information, storing it, using compute, and creating all sorts of dashboards and things like that. I wouldn't have thought that was how we would have scaled up so quickly.
Mm-hmm.
It's been a great surprise, and the demand is very strong across the install base as well as new customers coming to us. Now that we've started with the agentic web, we're actually getting a lot of interesting inbounds. You know, it sort of reminds me of the early days of Akamai when you'd walk in.
Mm-hmm
... when people used to have desk phones, and you'd see the red light on. You know, people would be calling saying, "I need, you know, need this stuff." We're starting to see that kind of activity again, which is kinda neat.
As I see investors sort of reengage with the Akamai story, and maybe they took a break for a couple of years, but kind of looking at the trend lines of the business, the question I often get when it comes to the public cloud business is: What gives Akamai or why do customers choose Linode versus the hyperscalers versus, you know, some of the altered clouds out there? I think one of the important points that you guys make is that all three major hyperscalers are customers.
Exactly. Yeah
of Akamai. If you just talk to the value proposition of Linode in the public cloud market.
You know, I think there's a, the misnomer out there that it's a zero-sum game. Most companies are doing multi-cloud, and there's applications, even with some of the hyperscalers, where we just get better performance, right? There's an application where you're just better served, whether it could be economics, it could be latency, where using a more distributed cloud is a better answer. You know, we've had companies trying to comply with standards in finance where they were complying with Apple Pay.
Mm-hmm.
You can't get that out of a, out of a public cloud, but you can get that by using us. We have some customers that might have workloads that are in a third-party cloud or a public cloud that their data is accessed very frequently. You store the data, but you're reaching in and grabbing that data, it's going all over the place. That's very expensive.
Yeah.
You have a lot of hidden costs. Egress fees, for example, are something that we've had customers say, "If I move to you, I'll use your compute and storage, and you're not gonna charge me for the egress fees.
Mm-hmm.
The reason we can do that is we have one of the largest backbones in the world, and we deliver hundreds of terabits per second, so the cost for us is virtually nothing. We pass that on to our customers, and we can differentiate from an economic perspective. It could be economics, performance. Sometimes it's just for diversity, where you wanna have multiple cloud providers.
Mm-hmm.
It could be maybe a major AWS shop, but you have two or three workloads with us. It's not necessarily a zero-sum game where one cloud takes everything.
Just a couple follow-ups there. In terms of those lower egress fees-
Mm-hmm
... can you give us a range on how much cheaper you are versus your versus the hyperscalers?
Mm-hmm.
The other question was around, Well, let me just stick with that first.
Sure. Yeah. A lot of times, even in the CDN business, we started to see this, and we actually started to build some products to help with our customers, where they might have their web infrastructure at a hyperscaler, and you have the notion of your origin, right? Where all the, say, if you're a video customer or even a retailer where you have lots of images and that sort of stuff. The idea of having a CDN, obviously, is you wanna get great offload, meaning every time a request comes in, you want it to hit an edge server and not have to go back to the origin, right? Better performance, lower cost. We have customers where sometimes we'll have 99% offload or 99.5% offload, and their bill for egress could be equal to what they're spending in CDN.
Mm-hmm.
For just that little bit of a cache miss, because it's so expensive. I don't have all the numbers handy, but you can look at some of the public costs for what it would cost for egress, and it's, you know, compared to CDN pricing, more expensive.
With the Linode Kubernetes Engine that you guys
Mm-hmm
... have launched, is that expanding, the types of workloads that customers can run? Is that becoming?
Yes
... a criteria for why customers gravitate to Linode?
Yeah, I mean, that obviously, like, if you think going back to the beginning of the discussion where our first origins were in this notion of functions as a service, you know, programming your cache, programming, you know, some basic content awareness where you could do, you know, tailor personalized content based on an IP address. You can program the CDN to do that. That's very basic stuff. Having a Managed Container Service, now I can run all sorts of different code and workloads-
Mm-hmm.
in that container that I couldn't do in the earlier days on the programmable CDN, if you will.
In terms of the hyperscalers, we've mentioned that, you know, they're customers of Akamai and their public cloud service. What are the hyperscalers using you guys for?
Yeah. We've talked sort of at a high level. Some are using us for video delivery, some are using us for API management, some are using us for advertising, decisioning, and things like that.
Awesome. Let's move the conversation to the inference opportunity.
Mm-hmm.
Akamai Inference Cloud was announced at the end of October, beginning of November. The edge computing opportunity has been discussed about as a potentially big market for several years, kind of across the industry.
Sure.
I would say it's generally been slow to ramp.
Mm-hmm.
maybe it's having its moments now.
Yeah.
With the opportunity around inference, what gives the you guys the confidence that a significant portion of AI inference will happen at the edge, and what will be the drivers for this opportunity to materialize?
Yeah, good question. you know, if you think about sort of the next iteration of applications, you know, today you might go to a retailer or to, like, a Home Depot or Lowe's where you interact with it with a search, right? You do a search, you get a bunch of stuff, and it's helpful. Imagine now you work with a virtual travel agent or a virtual contractor. The amount of data that needs to be processed and the amount of compute that happens is much more robust, right? You might have a small language model that you're using in real time, and you also wanna make sure that that is, you know, trained up and informed, etc .
The amount of compute for that is pretty significant, and you couldn't do that using, say, like, our EdgeWorkers or, you know, program that in JavaScript and that sort of stuff. Latency is a big issue, right, where just like in the old days when you had a shopping cart, if your shopping cart took forever to load and you see the circle going around, you're gonna abandon that. You know, either leave the site or refresh it, and the vendor, you know, the partner of ours, our customer will lose a sale.
Mm.
Same thing will happen in this world. It's just now compute is gonna be a much bigger component, right?
Mm-hmm.
In order to deliver that experience, maybe that virtual travel agent is, you know, booking your dinner reservations for you. It's putting markers in your calendar for, you know, travel. It might be getting you a car service and recommending all sorts of different things. That's a much better experience, but it requires a lot more compute.
Mm-hmm.
Latency is a big issue, right? workloads like robotics, you need to be close to where the factory is. Autonomous driving is gonna require extremely low latency.
Mm.
Those are just a few ideas of what we're seeing from customers. What's interesting is when we started working with NVIDIA and doing the trial to see what can we get out of these, you know, RTX 6000 PROs.
Mm-hmm.
What is sort of that, you know, token to megawatt mathematics look like? We were blown away with what we can do with those in even relatively small deployments. Say maybe, you know, a megawatt of power gets you about 1,000 GPU.
Mm-hmm.
You can do some pretty powerful workloads. You know, we have customers coming to us, anything from, say, AI startups that wanna do some rentals and just get access to GPUs. We also have others that wanna do clusters, where they say, "I've got a Maybe it's a research project I'm doing. Maybe it's a post-training. Maybe it's robotics. Maybe it's a really robust advertising decisioning model that wants to work on the fly and requires an enormous amount of data.
Mm.
We're starting to see some really interesting use cases now and a pretty nice pipeline that's building up, and latency is a factor. As a matter of fact, with this very large customer, deployment.
Mm-hmm.
The latency was the requirement here was such that typically within the U.S., you get pretty good latency. Like, let's say if you're in the middle of the country and you're on 1 of the coasts, it's usually tolerable.
Mm-hmm.
In this particular case, it wasn't. We had to be very, very close to where the application was being run.
Yeah. That makes a lot of sense. You mentioned that that large customer, it's a $200 million 4-year deal with a major tech customer at the forefront of AI. Why did that customer choose Akamai? What needs to get done between contract signing and the timing of revenue, which you sort of pointed to.
Yep.
At the, you know, at the second half of the year, maybe Q4 in terms of where things are. Just the timeline between, well, why did they choose you guys?
Yeah.
The timeline between to get revenue to start to materialize.
This, this was an existing customer, not one of our top 1% customers. It's been a customer for a while. When we announced the Inference Cloud back in the fall, you know, they had called us and said, "Hey, we've got this unique circumstance where we wanna run a cluster of GPUs, and here's the type of performance we're looking for. We wanna trial it with you." Right? We actually have some like NVIDIA does refer some business our way too, where customers will come to them and say, "Hey, I've got this use case. I wanna do X, Y, and Z. I need this type of performance. How would I go about doing it?" In some cases, working with us is the best chance you have to meet certain requirements.
It was just a, you know, an idea at first, and then it was, "Let's run a trial." You do a proof of concept, and you run that on a small scale in, you know, one of our locations and see what type of performance you get. Once the customer was comfortable with that, they placed a very large order. In this case, it's a fairly good-sized cluster. We had to, you know, light up some more data center space, in particular for this.
Mm-hmm.
In terms of timing, signing a contract of that size, it's an existing customer, so that helps. You usually have an MSA in place and that sort of thing. It does take, you know, time, approvals internally and that sort of stuff. There might be three to six months to do a contract from proof of concept to when it actually gets signed with that scale. Maybe it's two months on the short end. There's a lag between, in this particular case, since we had to get some new space before we get revenue.
Mm.
The pipeline is informed from existing customers, new customers, referrals from our channel partners, also from some of the partners we have on the technology side, like NVIDIA.
Yeah, it's great context. I wanna have a discussion on what it takes for the build-out of the AI inference opportunity.
Yeah.
We have about 20 locations today supporting Akamai Inference Cloud. What does the build-out of co-location facilities look like over the next 12 months, and then maybe looking beyond that?
In, in some of the-- Like, right now, the 20 locations we have were all existing facilities, and we just can roll out racks of GPUs. All of the sites we were in, you know, had all the necessary cooling requirements and things like that, so that wasn't a problem. With certain clusters, if someone says to us, "Hey, I want a 1,000 GPU," we may not have space in the particular areas, we need to add on to an existing site or maybe deploy into a new site. There'll be a mix of rolling out into existing sites where we have capacity. A lot of our-- The way we're building our CPU or our compute centers is essentially you look for our bigger sites, the 41 sites or so, are generally 5 MW, typically the initial size.
I may only be using one or 2 MW, and I'll roll up to five, but I'll usually have an option to go to five, 10 or 15, depending on, you know, the type of relationship. A lot of the folks we work with will be doing a multi-year build-out. There might be campus 1 is available today, and then we have an option on campus two that might be available next year. There'll be continuous build-out of capacity. In terms of, like, if you wanna think about unit economics, I've given this statistic, I did a little bit of this on the call. If you think about just GPUs in general, for a 1,000 GPU, we'll stay in base 10, it makes it easy.
Mm-hmm.
It requires about a MW of power. It's about 600 kilowatts just for the GPU, but with all the other stuff, it's about called a MW of power.
Mm-hmm.
Okay, a MW of power in the US, you can get it for anywhere from $2 million-$4 million a year, typically.
Mm-hmm.
California is a little bit more expensive, maybe $6 million. You know, generally speaking, somewhere in that range. For the CapEx, I think it's funny, I did this the other day. I like to find public references. I went to ChatGPT and said, "How much would a 1,000 GPUs cost with all the servers, the switches, and all this other stuff?" It gives you an interesting answer. It comes up with anywhere from, say, $12 million-$16 million. Let's round that up. We'll call it 20.
Yeah.
On the high side. The rental market today is around, I think we launched on Saturday, $2.50 an hour. Okay, $2.50 an hour for 1,000 GPUs for the year would be $22 million, roughly.
Mm-hmm.
At $2, it's around 17.5. Okay? Let's take that number and divide it in half even.
Mm-hmm.
Okay? If you think about the unit economics, I've got high-end $20 million of CapEx. Depreciate that over six years, that's roughly $3 million, $3 and a half million.
Mm
... of depreciation. Couple, $2 million-$3 million of co-location costs. Even if I get half of that price, it's still very good unit economics for me. The operating margin is in the 60%, 40% range or better.
Yeah.
At two, at $2 is phenomenal.
Yeah. I had a whole section on unit economics, but you got in front of me, and that's super interesting. In terms of, on the theme of the build-out of the AI inference opportunity, how do we think about the existing network?
Mm-hmm
... and growing to that? whether it's the roughly 50 Linode regions.
Mm-hmm
... that you have, the broader 4,000 plus-
Yep
... points of presence, is that going to be leveraged in terms of this build-out, or are you building out like net new, or renting out?
Yeah.
Net new co-lo.
Yeah. There'll be a combination. I'd say, the hierarchical compute capabilities functions as a service available in all of our servers.
Mm-hmm.
That's sort of serverless request-based.
Mm-hmm
... edge Java, you know, WebAssembly type programming for, you know, basic things.
Mm-hmm.
The container service we actually have available in all of our locations as well.
Mm-hmm.
Limited amount of capacity, obviously. Some of our locations are pretty robust, where you might have, we call them our multi-link E-cores, is the term we use internally, which effectively is a big server farm where you have most of your big telcos all connecting at once. You've got maybe several hundred to maybe 1,000 servers or something like that in those locations. There you've got more power, more space, where you could run your GPUs in those locations. We could leverage those and roll in some racks of servers and get, you know, maybe 100 locations or something like that.
Mm-hmm.
There will be purpose-built locations for customers when the math makes sense, where they have, say, a cluster that they need in a particular area, we'll build purpose-built for that. It'll be a mix of leveraging some existing sites, but some of the sites are like in telco hotels.
Mm-hmm
... where you don't have the power and the cooling necessary to run even a compute server. The 1U, 2U rack mount servers you use for CDN, pretty cheap. They don't require a lot of power, don't require a lot of cooling.
Mm-hmm.
You can run those in, you know, pretty small locations. But even the compute servers are much more robust for CPU that you wouldn't run them there. Also, I don't know that the economics would necessarily make sense for somebody to run GPUs in 4,000 locations.
Mm-hmm.
The math probably wouldn't work at that point. It's not saying that we couldn't go a lot more distributed than, say, 100, but practically speaking, probably 100 locations for GPUs at some point makes sense.
Mm-hmm.
You know, we'll probably grow from 20 today to somewhere 20-40, something like that. A lot of this could be informed by, you know, customer demand, will be the main driver of why we pick additional locations.
Just as a follow-up that I think the team mentioned that the initial capacity for the 20 has already been contracted out on the, on the 20 existing locations. How do we think about where do we stand in terms of the utilization potential and potentially going denser in, in those 20?
Yeah. We didn't spend a heck of a lot of money there. There wasn't a huge revenue opportunity.
Mm-hmm.
You know, so.
Just sort of proving out the concept.
Yeah. It was proving the concept out. We do have a sliver of that that we use for proof of concept. You know, like I said, the typical sales motion is someone comes to you and says, "I wanna run a proof of concept." The proof of concept might be running in a couple locations using a few GPUs or whatever. That we do on a regular basis. We do have some of that, you know, in terms of being sold out. That is set aside for proof of concepts, but it wasn't a huge revenue opportunity. The bigger revenue opportunity is this initial buy of $250 million of CapEx. Only a portion of that goes to this big customer. There's still thousands of GPUs to sell there.
There's an enormous amount of revenue to come, with that as well.
It's also the basis of my next follow-up. Is the timing of these purchases on the GPU, CPU side, does that happen once a contract has been signed, or are you purchasing ahead of that, anticipating future demand?
I mean, I would love to be in the situation I was just in where I had a big customer willing to... You know, if I, if I didn't have the chip inflation, I probably would've covered practically all my CapEx, right?
Mm-hmm.
That's an awesome position to be in.
Mm-hmm.
It will be informed by the pipeline, growth of existing customers and demand. You know, we will invest ahead of time like we do. It will be informed based on what we're seeing in the pipeline and what customers are doing.
Awesome. I wanna talk, spend the next couple of minutes just talking about some of the debates coming out of Q4, which, I mean, the results in Q4 were really strong. You have a CAS business that's accelerating, global cloud business accelerating, $200 million customer inference service. One of the things, though, is that margins did come down to 26%-28% from about 29% in 2025 as CapEx steps up, I think to the range you provided was 23%-26% versus 19% of revenue in 2025. With the full understanding that growth doesn't come free.
No.
We need to fund this investment, how should we think about the margin trajectory from here as the Inference Cloud business scales?
Yeah. If you look at, we gave this disclosure on the call. Where we're seeing the pressure points on margin are in cost of goods sold primarily around co-location-
Mm-hmm.
Then depreciation. If you look at that, almost the entire amount is those 2 items. With co-location, there's 2 factors. 1 is we are buying ahead of time. As I deploy these thousands of GPUs, with this latest purchase, well, I haven't sold them yet, so I don't have any revenue against it. Even with this big customer, I'm gonna have that facility up and running for several months before it ramps fully up with revenue. There's the timing aspect, and that goes for my entire compute platform, right? I've got capacity. Anytime I have capacity available, I have some inefficiency in my math on co-lo.
Mm-hmm.
Which you'll always have that. The other thing is, as I talked about how we buy some of the bigger locations, we'll work with our vendors for 7-year deals, 10-year deals in some cases. There you're making a fairly good-sized financial commitment saying, "Okay, I know I'm gonna be using, you know, 5 MW over the 10 years, and I want expandability, but I want unit economics today at that purchasing power." I might make a, say, a $50 million commitment over a 10-year period, and maybe I'm only spending $1 million in cash.
Mm-hmm.
with the lovely lease accounting standard.
Yeah.
You need to put an asset and a liability up, and whenever you have a lease that has either free rent or has escalations or commitments, you need to straight line that.
Mm-hmm.
There is some element of non-cash co-location costs in my cost of goods sold. There's some inefficiency there. That's part of what's driving it. That said, there's a trade-off between growth and margin. If I wanna grow faster, let's say a customer came to me and said, "Ed, I wanna spend $500 million with you." Whatever, pick a number.
Mm-hmm.
Well, my margins are gonna go down a touch before because my payback on my servers might be two years or whatever. I have to get co-location for if there's a deal like that, I don't just have that sitting around necessarily unless I wanna distribute it everywhere in the perfect locations and find everywhere that I have capacity. That's not reality. It's probably gonna be in, you know, several locations. It might be a U.S. opportunity or a Europe opportunity or whatever. There's gonna be some inefficiency associated with that. Where we get good operating leverage is on the go-to-market side, on the engineering side, and on the operations side.
Mm-hmm.
You know, we have some overlay sales for compute. It's a relatively small group compared to the size of our sales force. The people that are going out and getting the co-location deals and managing the operations associated with that, doing the deployments, buying and, you know, testing the hardware and the different routers and switches and chips and all that stuff, same group that does that for CDN. CDN basically runs our security. Security runs on the same platform. I get an enormous amount of leverage there. We've moved about 1,000 engineers out of our CDN business into our compute business because there's a lot of similar things that you do-
Sure.
in terms of doing your engineering and running the distributed platform. We will get a lot more leverage on the operations side over time.
Maybe if I look at the unit economics questions from a through the lens of a customer, so sort of lifetime value of a customer. Let's say it's maybe not this particular $200 million customer, but if you sign similar sized deals and with commitments for four to five year ranges, what do you expect the contribution margin to look like over the lifetime of that customer commitment?
Yeah.
I was sort of referring to, I think a couple quarters ago, you mentioned that you think that maybe the inference cloud business could get back to, kind of where your public cloud gross margins are.
Yeah, I mean, I just if you go through the numbers we just talked about, let's say you get per thousand, let's say you're getting $12 million-$15 million of revenue.
Mm-hmm.
Against costs of say $3 million-$4 million for co-location to, you know, $3 million of depreciation. Your operating margin is north of 30%.
Mm-hmm.
Your gross margins are in the 70%-ish. You can get there, certainly. That's, you know, assuming a discount off of what the published price for rentals are.
Yeah.
The other misnomer I think that's out there is people believe, well, six years depreciation, that's probably not the right way to think about it, because what about the next generation? Well, sure, there'll be a next generation, and in that six years, probably many next generations. That doesn't necessarily mean that everyone's gonna move to that next generation.
Mm-hmm.
Like, one of the things I'm finding is customers will come to us and say, "My particular application runs best on this chipset." Maybe it's AMD, maybe it's NVIDIA, maybe it's Intel. I also have a particular workload and a set of economics that I might not be able to afford the next generation of chips. It might not make sense.
Mm-hmm.
I look to our Linode business. Before we bought Linode, they offered GPU as a service.
Mm-hmm.
It's been a nice little business for them, and we're selling, you know, generations that were put in place before we even bought Linode. We've owned Linode now for four years.
Mm-hmm.
I anticipate, I mean, I could be wrong on this, but I don't think I am. There will be a market. Maybe the unit prices come down a little bit.
Mm-hmm.
I get cost, good cost recovery of say one to two years on my servers. The same is true with CPU. You know, it's kind of that dollar CapEx, dollar revenue gets you about, you know, one year or so.
Mm-hmm.
Even if it's $0.85, you still get somewhere between a year or two of, you know, maybe it's 18 months of recovery on your CapEx. It's, it's a, it's a great business. I'm not too concerned about, you know, Most customers will probably renew.
Mm-hmm.
You know, unless their application goes away. Generally speaking, it's tied to a business that has some notion of scale with it as well.
Mm-hmm.
That you might say, "Hey, I wanna start off with X amount because I'm doing, you know, an application that is tied to something else I'm doing that's revenue generating, and as long as that's growing, I'm gonna continue to need more and more and more." We do see that. We see very nice, call it same store sales growth of folks that are using it. That's very true with the public clouds.
Maybe my last question, because again, time about security in the delivery business, so I'll kind of bundle the question. You're confident in sustaining the growth that you have been posing?
Mm-hmm.
In security, you know, I think it was 9% constant currency.
Yep.
I think you guided to high single digits.
Mm-hmm.
You're confident in sustaining that level of growth in security. In the delivery side of the equation, it does seem like the pricing environment's a little better.
Yes.
Fewer competitors out there. We got Olympics, and we got the World Cup.
Yeah.
how should we think about the growth trajectory in delivery-
Yeah.
Over the next year?
Sure. Delivery has gotten a bit better. One of the things that's a little difficult to digest this year is we had acquired the assets of Edgio. Throughout the year, we had modeled it that some of these customers, some of the bigger customers, were multi CDN, and we bought the other CDN.
Mm-hmm.
They have to rebalance. Let's say we have somebody who's doing 50% with us, 50% with Edgio. We combine it, we probably end up with 75% share as a general statement. 25 goes to find somewhere else. That happens throughout the year. It's not all like I closed on December 20th and December 21st, everything rebalances. It happens gradually throughout the year. You did see revenue declining.
Mm-hmm.
Part of that was, like modeled and planned. We knew that was gonna happen, that someone else was gonna pick up some share where they didn't have share before. Now, we are seeing the pricing environment get better, and we're taking a much tougher stance with pricing now saying, "Look, we just talked about on our fourth quarter call, $200 million of additional CapEx just because memory prices went up.
Mm-hmm.
Right? Co-location is not getting cheaper. It's, you know, some markets it's solid, but other markets more expensive. Labor doesn't get cheaper. We can no longer eat all of this. We have to start passing along some of that price, and certainly not accepting a 15% or 20% price decline.
Mm-hmm.
In the delivery business. We are taking a much tougher stance. We are trying to raise price on renewal. We'll be successful in some areas, on others, we may not be.
Mm-hmm.
That's a much different stance from us. We've never raised price before. We're doing it for the first time in our history. On security, I gave out some really interesting statistics on the fast-growing parts of the business. API Security exit over a $100 million run rate, growing at over 100%. The penetration rate in the install base is sub 10%. Very applicable for our WAF customers to use API Security.
Mm-hmm.
That's got an enormous runway. I actually think in the agentic web, as we talked about earlier, with all the data flows that go with that more robust application, that's gonna be a massive attack vector. We think there could be even a second generation of growth out of API Security protecting those agentic applications in the future. In terms of Guardicore, again, sub 10% penetration of the base. The majority of that business is coming from new customers. Very robust channel there. Very great dynamics with existing customers. The typical motion is to add more agents to protect a greater slew of your network. You start off running maybe a quarter of your network, and then you expand our license business, which isn't a significant business, but has 90x% renewal rates.
Mm-hmm.
Usually it's an upsell. They're term licenses, so very healthy dynamics in that part of the security business that was growing, you know, at over 30% last year.
Well, we're out of time. Ed, thank you so much for giving the update on the Akamai story. A lot of great things happening in the business, and all the best for 2026.
All right. Thank you, Sanjit.
Thank you.
Appreciate it.