Good day and welcome to the Second Quarter 2024 Akamai Technologies Earnings Conference Call. All participants will be in listen-only mode. [Operator Instructions]. After today's presentation, there will be an opportunity to ask questions. [Operator instructions]. Please note this event is being recorded. I would now like to turn the conference over to Mark Stoutenberg, Head of Investor Relations. Please go ahead.

Thank you, operator. Good afternoon, everyone, and thank you for joining Akamai's second quarter 2024 earnings call. Speaking today will be Tom Leighton, Akamai's Chief Executive Officer and Ed McGowan, Akamai's Chief Financial Officer. Please note that today's comments include forward-looking statements, including statements regarding revenue and earnings guidance. These forward-looking statements are subject to risks and uncertainties and involve a number of factors that could cause actual results to differ materially from those expressed or implied by such statements. The factors include any impact from macroeconomic trends, the integration of any acquisitions, and any impact from geopolitical developments. Additional information concerning these factors is contained in Akamai's filings with the SEC, including our Annual Report on Form 10-K and our Quarterly Reports on Form 10-Q. The forward-looking statements included in this call represent the company's view on August 8, 2024. Akamai disclaims any obligation to update these statements to reflect new information, future events, or circumstances, except as required by law. As a reminder, we will be referring to certain non-GAAP financial metrics today. A detailed reconciliation of GAAP and non-GAAP metrics can be found under the financial portion of the Investor Relations section of akamai.com. I will now hand the call off to our Co-Founder and CEO, Dr. Tom Leighton.

Thanks Mark. I'm pleased to report that in the second quarter, Akamai delivered continued strong momentum in compute, strong growth in our security portfolio, steady operating margins and healthy earnings growth on the bottom-line. Second quarter revenue grew to $980 million, up 5% year-over-year as reported and up 6% in constant currency. Non-GAAP operating margin was 29%. Non-GAAP earnings per share was $1.58, up 6% year-over-year and up 9% in constant currency. These results were in line with or above our guidance. Before I provide more color on our performance, I'd like to review how Akamai is evolving as we grow. As most of Akamai first made its name with the invention of content delivery services, and we're still the world's leader in that market today. We stand out for providing the scale and performance required by the world's top brands as we help them deliver reliable, secure and near flawless digital experiences. Recent examples include delivering the Euros football tournament and the summer games in Paris for top broadcasters around the world. As we've said in previous calls, our delivery business has been challenged in recent quarters by macroeconomic and geopolitical headwinds. Our plan for delivery is threefold. First, we will remain disciplined when it comes to the profitability of traffic that we choose to serve. Second, we will continue to leverage our market leadership position and installed base of major enterprises to generate cross selling opportunities. And third, we will continue to take steps to retain our market leadership while also reinvesting most of the cash flow from our delivery product line into the fast growing areas of the business. In Q2, delivery accounted for one third of our revenue, or $329 million. This is quite a change from five years ago when delivery accounted for two thirds…

Thank you Tom. Today I plan to review our Q2 results and then provide some color on our expectations for Q3 in the full year. Turning to our second quarter results, total revenue for the second quarter was $980 million, up 5% year-over-year as reported, and 6% in constant currency. Compute revenue was $151 million, up 23% year-over-year as reported, and 24% in constant currency. We continue to be very pleased with the level of enthusiasm in the market as more and more customers are leveraging our enterprise compute solutions. In particular, we are seeing a broad array of enterprise compute use cases including live transcoding, secure access, observability, object storage, real time log aggregation and insight spatial computing and deep learning AI models across many verticals including media, e-commerce, software and financial services. Moving to security revenue in the second quarter, security revenue was $499 million, up 15% year-over-year as reported, and 16% in constant currency. We're very pleased by the continued performance of our Guardicore Zero trust solution and highly encouraged by the traction we are seeing in our recently launched API security solution. It's worth noting that the no name transaction closed in late June and the revenue contribution in the second quarter was less than $1 million. Combined, compute and security revenue grew 17% year-over-year as reported, and 18% in constant currency, representing 66% of total revenue. Moving to delivery revenue was $329 million, which declined 13% year-over-year as reported, and 12% in constant currency. The decline in delivery revenue was primarily related to the revenue impacting items that I outlined last quarter and was in line with our expectations. International revenue was $471 million, up 3% year-over-year and up 5% in constant currency, representing 48% of total revenue in Q2, foreign exchange fluctuations had a…

We will now begin the question-and-answer session. [Operator instructions]. The first question today comes from Patrick Colville with Scotiabank. Please go ahead.

Thank you so much for taking my question. Frank and Ed, really great to be part of the hack. My story I want to talk about the business makes shift. I mean, that's where you opened your prepared remarks. Specifically, I want to focus on compute. The $100 million revenue you just called out from Akamai connected cloud is really compelling and great to see that ramping. When might that hockey stick to become an even greater revenue base? What's the trajectory of Akamai connected cloud over the coming quarters? And if we think out beyond that?

Yes, great question. And I got to say, we were very pleased to see the rapid early adoption. that's a capability that we really just started selling in earnest this year. We had some very early adopters towards the end of last year. And, if we can get up to $100 million ARR by the end of the year, which we think we're going to do, that's great, for the first year of the product, and then we'll see where we are at the beginning. Next year, we'll give guidance in February for the year in compute, but we're very optimistic about strong growth in compute driven by the enterprise customers and our new capability. There's an enormous market there, obviously, and so we're really looking forward into tapping into that.

Very helpful. Thank you. I guess the second part of my question I want to ask about delivery this year. You've been very clear about the headwinds to the delivery business in 2024. I appreciate you might not want to give guidance beyond 2024, wondering whether the headwinds we're seeing right now are cyclical headwinds or are they structural in nature? Thank you.

Yes, I don't think what we're seeing today persists over the long term. traffic, I think, will grow, continue to grow maybe at a little bit of a slower rate, than it did certainly during the COVID times. But, delivery, I believe, is here to stay. We are intent on remaining the market leader by a good margin. It's a very profitable business for us. We're very careful about that. We're very efficient in what we do. And it's very synergistic, with our security web app firewall business and our emerging compute business. So I don't see these headwinds persisting over the long term. There are geopolitical considerations that we're worried about, for next year. But I don't think this is a long-term phenomenon. And in any case, given the very fast growth of our security and compute product lines, they've nearly trickled in revenue over the last five years. So now where they're two thirds of our revenue overall, I think the, what you see with delivery, with sometimes challenges, sometimes good, it has a lot less impact on the overall growth rate as we go forward. Ed, do you have anything you want to add to that?

No, I think you covered it well.

The next question comes from Keith Weiss with Morgan family. Please go ahead.

Excellent. Thank you guys for taking the question and congratulations on the solid quarter. And I wanted to ask you a little bit about kind of parsing out the guidance, particularly the full year guidance. If I'm doing my math right, the midpoint of the full year comes up by about $5 million for the full year. But we're adding or no names. It sounds like about $20 million in revenues for the full year. Is there a part of the equation that's coming down, a part of the business that you're getting more cautious on? That makes up that difference?

No, Keith? Yes. So we included no name in our guidance last quarter as well. So there's nothing that's changed. If anything, the business has gotten a little bit better. So our guidance has come up a bit to reflect that.

Got it, got it. And then on the expense side of the equation, the savings from moving sort of in house from the hyperscaler is $100 million is real savings. Congratulations on that. That's quite a feat. Tom, you talked about the ability to sort of start pushing that more into operating margins in the near future. Can you give us an indication of what near future means? 2025 near future, or are we thinking two or three years out or further?

Well, Yes, the saving, operating savings we're getting, as Ed said, we've been plowing that back into the business by and large so that we can invest in growth. There's more savings to come there, but we really get the upward pressure on margins, the beneficial tailwinds on margins as the mix shift continues. As we add compute customers, that is good margin for us. And it's accretive security as we add customers there, it's accretive. Now, as Ed noted that today with the new security products, initially they're dilutive, but as we grow the revenue there, every customer we add, every deal we sign improves margins. So that over time, 30% is our goal, we're very close to that today, but over time, we think we have good potential to grow beyond that.

The next question comes from John DiFucci with Guggenheim Securities. Please go ahead.

Thank you. I have a question on Guardicore. So, segmentation broadly seems like it's becoming more relevant in the market. Customers are more accepting of it. It's no longer like a new thing, and it has been for a while, and you guys have been there, and you bought God of a couple of years ago, but frankly, it seems like an essential component to a zero trust environment. Not everybody has it. So can you talk a little bit more about how about this business, but really about your Akamai Guardicore Zero trust platform that you just launched a few months ago, and how that sort of fits in the ecosystem of some of an enterprise when they need to protect all their assets to establish that zero trust environment. Because it always seemed to me that this was essential, like I said, for zero trust. I guess if you can also, in addition to the technologies and what else it fits in with, can you also hit on your channel efforts regarding this platform? Because I know this is sold through the channel. I think, Ed, you said this before, but it seems like a really sophisticated solution. It's not just you're buying a firewall for somebody or something like that. If you can just talk a little bit about how. I know it's only, it's early, but how that's working through the channel?

Yes. I think you characterized it very well. Segmentation is essential, I mean you got to lock the doors in the windows as best you can, but now we're still gets in. And I think really the most important thing an enterprise can do is lock down everything inside and that means Guardicore. It means having your agent on every application on every device. And you're right, most enterprises don't have it. When you go back a few years, and I think the community of large really disfavored segmentation. And that's because the way it was done back then was really crude. You did it in hardware, it's very inflexible, hard to do. And at the end of the day, if you did it at all, you had giant segments which defeated the whole purpose. You weren't very secure because the malware would get in and wipe out an entire giant segment and you had a big problem. And Guardicore solved that problem through software, very easy to manipulate, make updates much more secure, fine green controls. And so it's been an education process for us in the marketplace. We viewed it as something that was going to be essential and that, I think, is proving out, as you noted. And of course, with the ransomware headlines and disasters, it's not surprising to see why people are waking up to what they really do need this. So now the next question is with the platform. And there, what we've done is combine the Guardicore, which is protecting inside app-to-app device-to-device communication with the employee device to internal applications. And so we've combined what's called North, South and East, West. Now again, you go back a few years ago, they were different buyers and treated differently. But then we were thinking…

And if I could, Tom, because that all makes a ton of sense to me. But it also raises the question, like, what are people -- what do they do? What are the alternatives? Like I'm familiar with Illumio and as another company, I've seen called Truxport but like you said, like people don't lot of enterprises don't even have segmentation implemented. So a lot of do, but a lot don't. And so what are the -- are there other things that we're just -- that I'm just missing like I don't know, is Palo buy the whole Palo platform. Are they just saying, you don't need it? Or we have something that kind of does it? Like I'm just trying to -- because the opportunity is just seems really big here?

No, I think it is a big opportunity. And very few, relatively speaking, enterprises have it today, the early adopters are the critical infrastructure companies because they really, really have to have it. And we do compete with Illumio. They're probably our leading competitor. We believe the Guardicore solution is a lot better. We actually have our own mini firewall in the Asia. We don't have to rely on the firewall in the OS, which sometimes won't be there, it might not be consistent. We can cover a lot of the legacy systems which that's important to enterprises to get more universal coverage. I think there is a ton of greenfield. And I wouldn't be there when one of the other competitors is talking about their platforms. They probably don't spend a lot of time talking about segmentation because they don't really have a solution for it.

The next question comes from Mark Murphy with JPMorgan. Please go ahead. Thank you very much.

Ed, what is your latest thinking on the FX headwind to the full year revenue forecast. I'm just curious if there's been any movement there from, I think, previously, you've been looking at that as, I believe, $40 million headwind.

Mark, yes, not much of a change. The dollar moved around quite a bit during the quarter, up and down, but it's pretty much exactly the same. So for the full year, it's about $40. I gave the guidance already in the quarter in terms of the impact quarter-over-quarter and for year-over-year, but it's still about the same, just around 40 for the full year.

Okay. And then, Tom, as a quick follow-up, you mentioned a pretty wide array of the workload types that you're seeing on your cloud computing platform. And you mentioned toward the end, deep learning and AI models. I'm wondering if you can double-click on that, for instance. What are the types of model? Are you seeing LLM, the text models or image models or something else? And is it possible to estimate what percentage of your cloud ARR might be relating to those newer types of AI workloads?

Yes. Great question. I would say today, AI workload is probably a small fraction of the ARR. I think over time, potential for growth there. As we talked about useful in security, applications, chatbots, tailoring content for commerce companies, ad targeting, recommendation engines. I would say that the models are smaller because they're more focused. The giant models sort of are used to learn everything, your chat GPT, you can ask it any question at all, have it try to be knowledgeable about everything. Those are giant models and we're not really targeting that business. But for our customer base, they tend to be a lot more focused on what they're trying to do. Maybe it's a commerce site, maybe it's an ad site, maybe it's a security company. And they don't need to learn the whole world to really provide value. In fact, we see that with our own solution with the Akamai Guardicore platform with the control interface, powered by Gen AI. Really, it's a very specific application, which means that you don't need the gigantic model to really provide the value. And that means it doesn't have to run on this giant suite of GPUs, it can run just great on our platform, which has GPUs but primarily CPU-based, which gives us much better ROI. And that works great for what our customers are looking to do in terms of their AI applications.

The next question comes from Madeline Brooks with Bank of America. Please go ahead.

Hi team, thanks for taking my questions. I want to continue on the discussion of Connected Cloud and you mentioned some nice wins in enterprise outside of your traditional CDN customers. But I guess I just wanted to double click on that and kind of get a little bit of color. What are those customers for nonmedia, non-e-commerce. What are the use cases that those customers are finding from Connected Cloud? And if you could just help us break out to growth of those customers maybe versus growth of your more traditional customers? Are they growing around the same in terms of their adoption of Connected Cloud?

Yes, we're seeing growth both within the base and outside the base. I think, for example, with our qualified compute partner program with observability, a lot of companies need that to know what's going on with their applications. Security companies would need that. Now we also have a lot of media customers. And I would say that's probably the biggest segment today. We have, by design, a full media workflow ecosystem now supported through our QCP program on the platform and so a lot of media customers starting to take advantage of that. Outside of that, for example, OS and firmware patch storage, personalized waiting room experience, improved page performance with hints and so forth. We talked about with AI tailoring the site for a user based on what they've been doing so far. Real-time log aggregation and insights into your logs, observability, kinds of things. We even have a PBX, a telephone switching system running on it, 5G Internet gateway running on it. So it does broaden the base which is, I think, exciting for us in the future. But today, probably the biggest segment would be media.

The next question comes from Alex Henderson with Needham. Please go ahead.

Great. First off, I think congratulations to an order on the great results out of both security and out of the compute. And I wanted to focus a little bit on the compute side of the business because I think, ultimately, that's the area that needs to be proven out to the Street more than anything else. Can you talk a little bit about the mechanics around what portion of the customer base that's converting to compute is coming from internal? What portion of the compute are true new customers? How many -- how much of the growth is coming from existing customers that are increasing their upsell? Just kind of look at it as if it was a traditional stand-alone business and give us some of those critical metrics that go into analyzing the success of that business?

Yes. I'll take our first half and turn it over to Ed. And the answer will be pretty similar to the last question. I would say our biggest users and the biggest segment for compute today is our large media customers. And that's by design. And our -- a lot of our QCP, our Quality Compute Partners are media workflow companies. So that's sort of the biggest segment. I would say observability as a capability, a very large one as well, and that spans across all verticals and would include new customers. So we do have a bunch of customers in nontraditional Akamai verticals that are using compute. And I think over the longer run, that opens up a whole new market verticals for us. But the biggest chunk today would be existing Akamai media companies is the biggest. And Ed, do you want to add some color on that?

Yes, sure. Alex, so as Tom mentioned, we're seeing growth in both existing applications for stuff that's been on the platform for a while, but the bulk of the growth is actually coming from new customers. The new customer additions is growing very, very quickly. We broke out some numbers for you last quarter, and that continues to ramp very nicely. And we're seeing a significant increase in the pipeline. We are seeing some new customers come to the platform. And what's interesting is we're probably seeing more workloads and repeatable workloads in areas even -- Tom talked about media, but outside of media. And we're seeing customers that may be relatively small CEM customers are fairly large compute customers. So I'd say it's across the board where we're seeing the growth, but it's mostly from adding new customers to the mix and then they start to ramp.

The next question comes from Fatima Boolani with Citi. Please go ahead

I wanted to be in on the delivery guidance and the expected performance in the back half. So appreciate you experienced a lot of the traffic degradation patterns in the first half. But I'm just curious why the trajectory of the business is actually worsening in the back half? And then I have a follow-up for Tom, please.

Yes, sure. So I talked a little bit about the expectations for Q4. Just based on what we're seeing now in terms of traffic growth and what we saw last year, we're not expecting the normal hockey stick to any significant degree like we've seen in prior years. And also keep in mind, we closed the transaction with StackPath and Lumen last year, and that's all delivery revenue. So that makes sure Q4 tougher comparison if you're looking at sort of year-over-year growth rates, that's going to skew your perspective a bit. And as we talked about on the last call, there were some dynamics going on with one of our larger social media customers. The good news there is we've got a good handle on that. That's sort of playing out as we expected. But as we talked about, those are the factors that as you put that into your model, why it may look like it's deteriorating a little bit. But I'd say the biggest issue is the fact that you're anniversary-ing the StackPath and move contribution from Q4 last year.

That makes sense. That's very fair. And Tom, for you, I think you've been very constructive around the compute opportunity. There are so many specific examples of the momentum you've been garnering in the compute franchise. But taking a step back as a broader strategy question for you, as you think about scaling that franchise and have it becoming an even bigger part of the overall revenue story. How are you straddling this notion of not using compute as -- or essentially ensuring compute ends up being wallet share accretive against your base as opposed to potentially managing a situation in which the delivery franchise sort of bumps along and compute sort of plugging the hole? Just what are some of the mechanisms you have in place to continue to drive the actual wallet share growth and accretion within existing delivery customers from where you are extracting a lot of net new compute demand for now?

Yes. Compute is different than delivery. So it's not a situation where delivery revenue was going into compute. That's not the case here. And that compute opportunity is orders of magnitude bigger than the delivery opportunity. And so I think, over time, it becomes a much bigger business than delivery. And I think delivery does its thing and it's a very good business for us in terms of cash generation, in terms of cross-selling. And in terms of actually of the economics of the platform, so we can go out there and offer compute at and especially for chatty applications and applications where data is moving around at a much lower price point than the hyperscalers because we have the delivery platform, but it doesn't -- it's not a situation where it's plugging a hole in delivery. I think compute is a huge revenue growth driver for us in the future, independent of anything in delivery.

The next question comes from Jim Fish with Piper Sandler. Please go ahead

This is Quinton on for Jim Fish. Maybe touching on that first question there. A competitor in the space recently talked about pricing pressures from some of the largest medium customers getting worse over the past couple of months. Are you seeing those pricing trends in the market that's maybe driving some of that second half weakness alongside, obviously, the StackPath and Lumen impact? Or are you not really seeing these pressures given your decision to move away from these kind of lower margin delivery opportunities?

Quinton, this is Ed. As we talked about, we had some large renewals this year. So obviously, those we've been through all those, and there is certain pricing pressure there. I'd say it's nothing different than what we've traditionally seen in the marketplace. I wouldn't say that it's significantly worse. I think the issue is just not as much traffic growth. So as you reprice a customer, you tend to see significant traffic growth. So the revenue declines don't persist as long as they have in a situation like this. So I wouldn't say if anything has changed in terms of the trajectory of the pricing. It's always been very competitive, it always will be. So that's really not the issue. The issue is just we're not seeing the type of traffic growth that we normally see.

Yes. That's really helpful. And then obviously, it's still really early here with the Noname acquisition. But any update you can provide on the integration between Noname and your kind of existing Neosec API opportunity? And maybe how you balance that go-to-market of those two platforms and how you can kind of leverage a full suite to kind of grow the wallet here within a customer.

Yes, we are now going to market with Noname. And as I mentioned, within 2 weeks of the close, we had a fully integrated with Akamai products, existing Akamai products, in particular, our web app firewall where a lot of the APIs would go through that. So I would say we're basically integrated today. We have some Neosec customers who we are maintaining over time, that will evolve into the Noname product with some of the capabilities from Neosec, so we get the best of both worlds.

Yes. Just to add the -- just add the Noname acquisition came with a pretty sophisticated channel as well as a number of specialists. So we have both of those, so that's going to help drive some sales. And actually, from the minute we announced the close to when we closed or announced the deal ultimately closed, we saw a nice pickup in deals closed. So not no impact on the funnel and the teams already out there selling. So very excited about that. I think we've just enhanced our go-to-market capability as part of the acquisition.

The next question comes from Jonathan Ho with William Blair & Company. Please go ahead.

Hi. Good afternoon. As you listen to customers and what they need or want from the compute side of things. Are there any core services or capabilities that you feel like you're missing or you're going to add pretty soon that are maybe potentially catalyst for even faster adoption?

I think probably the biggest difference today would be the size of the marketplace. Obviously, the hyperscalers have an enormous marketplace and we're getting started there. We're really excited that I think now we have a very competitive media workflow marketplace. I think we've got a very competitive observability marketplace. And that's something that we're going to be continuing to grow. We're also building out, as you know, our distributed compute capabilities so that we'll be in more locations than in many locations where the hyperscalers don't have a presence, which will give us an advantage in performance and also in countries where you've got data sovereignty laws. We'll be in a better position to handle that. But it's yes, it's ongoing. We're continuing to develop and improve the platform, including with storage, a lot of effort going on there too. So that's going to be -- that will be ongoing for the foreseeable future. But if you can tell from the results, we're in a position now, we can get out there and be selling it. And it's great to see the rapid early adoption.

That makes a ton of sense. Just in terms of the delivery business, as we continue to see this decline as a percentage of revenue, it seems to be carrying the business in terms of margins. How concerned are you over deleveraging effects and CapEx efficiency as we see sort of the two sides of the business move in opposite directions?

Yes, I said it's not a huge concern there. The margins of the new products are very high gross margin products, so that will be helpful. Sure the compute business is a bit more capital intensive. But we've been able to drive down the CapEx of the core business and delivery pretty dramatically. So you're down low single digits as a percentage of revenue for that business. That will maintain as long as the delivery, as long as traffic is not growing significantly. And part of our strategy and being more selective of the type of peak traffic to average that we're taking on the platform is by design, is to make sure that we do maintain that efficiency as we're spending more CapEx in the compute business, and we're not taking on that sort of not as profitable peak to average type traffic so that we can maintain a low CapEx posture in the delivery business.

The next question comes from Tim Horan with Oppenheimer. Please go ahead

Kind of a few part question on cloud. Can you use your own platform? It sounds like you're moving a lot of legacy services on there to create kind of, you think, unique services for yourself and new services or improve legacy services on that platform. And then secondly, if you look at Microsoft, Google Cloud flare, they're kind of growing cloud in the 30% range. Do you think you can kind of get there? And are you kind of maybe CapEx constrained to do that? Or just maybe talk about how you can kind of get up to your peers there. And what do you have to do in the SMB market to hit that type of 30% growth?

Yes. On the first question, we have built capabilities for ourselves as part of our migration. So we're off of Snowflake and Data bricks which we had big spends there. Looking at over time, making our capabilities available to customers, again, a service that is more efficient, which I think is really important for customers. In terms of the growth rate, I would say that you want to compare the enterprise compute number, which we've talked about was a $50 million ARR at the end of Q1. We think that will more than double by the end of the year. That's sort of the number you want to look at that's comparable. We've got more in the overall compute number, but those are -- have products like image and video manager, legacy Akamai net storage, other kinds of things. which aren't as comparable for what you're looking at in terms of the hyperscaler growth. So if you focus on the enterprise compute, which is really going to be the growth driver for us, that's going in a very fast percentage now and, of course, on a much, much smaller number than the hyperscalers. Ed, do you have -- want to add to that?

No, I think that's exactly right. And that's something that, over time, as that number becomes more material, we'll start to break that out for folks to make it easy for you to see where that growth is coming from. But Tom is absolutely right. That's where the big market is. That's where we're seeing the explosive growth, and we see that we think that can continue. Obviously, the pipeline is growing. We're seeing a lot more use cases than we expected. I see a great participation from all of our reps across the world. It's not just one geo. So we're very excited about it.

And do you think you're capital constrained at all or product constrained at all in the enterprise there?

No, I don't think there's a capital constraint problem. I mean you could envision perhaps as a customer that may come to us with a big task where you may have to do some build out if it's in a particular concentrated geography. But we got a very strong balance sheet. We produce a ton of free cash flow. So there's no issue from a capital constraint perspective. I think we can grow this business to a significant size over time, and I don't think capital is a problem right now.

The next question comes from Rudy Kessinger with D.A. Davidson. Please go ahead.

Hey, great. Thanks for taking my questions on security. I guess in the second half, if I look at it adjusting out Noname, it looks like organic growth at constant currency is just about 11% to 12%. Obviously, it's been a slowdown versus the last several quarters. And in general, the security growth rate has kind of been pretty volatile over the last 5 to 6 quarters. Could you just give us the puts and takes on maybe some tough compares in the second half. I think you had some spike in DDoS strength last year. But also just going forward, just what's the kind of right growth range that we should expect out of the security business?

Yes, sure. So just in terms of some of the puts and takes, just remember, last year, in Q3, we had little over $6 million of license revenue, that's a couple of percentage points. So that's going to make your Q3 compare a little bit more challenging. The other thing to keep in mind, too, last year, we introduced some new security bundles for web app firewall that did phenomenally well. So we've anniversaried that, so that makes sure our compare is a little bit more challenging. And then between Guardicore and API security as they start to ramp, we think APIs going to ramp very, very quickly. It's just a smaller number. So as you've talked about, the growth has bounced around a little bit over time. That happens as you bring new products into the market and they start to ramp up, think about when we brought [ Batman ] manager to the market. It was a small product and then go to hundreds of millions of dollars. I think the same thing you'll see with Guardicore and API security.

Yes. Okay. And then on the delivery outlook, I guess, your -- I mean, fastly, I'll say it more directly, I mean, they certainly seem to indicate that outside, it was broader than just one social media customer has several large media customers that seem to be shifting traffic to lower-cost providers. It sounds like you guys aren't really seeing that dynamic or maybe it's not the year and it was already factored into the guide. Any comment on that?

Yes. No, we're not seeing a phenomenon of someone moving to low-cost providers. As a matter of fact, there's two less of them in the market today. So we're not seeing that. As I talked about earlier, we have a tough compare with a [indiscernible] Q4. And then we've just seen just a lower traffic year. Gaming has been unusually weak, video traffic isn't as robust as it normally is. That stuff happens from time to time, but we're not seeing a shift to low-cost providers or any new low-cost providers in the market.

Okay. The last question today comes from William Power with Baird.

This is Yan Samilton for Will. Ed, just going back to that more muted Q4 seasonality you're expecting for delivery again this year. If I remember correctly, last year, you were pointing to weaker trends in retail, including an uptick in bankruptcies there and then also weaker in terms of gaming. And now I know it's probably still a little early to forecast though, but is it those same verticals where you're expecting weaker traffic again this year that you're informing your expectation? Or is it some others or maybe it's more broad-based?

Yes. I'd say it's a combination of those two verticals. We've seen sort of over time the retail seasonal or to be less and less. Some of that has to do with the Zero overage product that we offer in the market. But just in general, it just hasn't been as robust as have been, say, 4 or 5 years ago. And then from gaming, yes, still weak. I haven't seen any major launches or we're not hearing anything from our customers that we leave me to believe that Q4 will be strong from that perspective. And then also, as I just talked about several times here, just traffic in general has been a bit sluggish from a growth perspective in general. So I don't see anything that tells me that, that's going to change going into Q4. So obviously, we've got a few months to go here before we get there. We'll update you when we talk again in November. But based on what I'm seeing today, just out of this worthwhile calling that out to folks as they build out their models.

This concludes our question-and-answer session. I would like to turn the conference back over to Mark Stoutenberg for closing remarks.

Thank you, everyone. In closing, we will be presenting at several investor conferences throughout the rest of the quarter and the rest of the year. We look forward to seeing you with those. We hope everyone has a nice evening tonight. Operator, you may now end the call.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.