Akamai Technologies Inc (AKAM) — Q3 2023 Earnings Call Transcript

Thank you, operator. Good afternoon, everyone, and thank you for joining Akamai's third quarter 2023 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 quarterly reports on Form 10-Q. The forward-looking statements included in this call represent the company's view on November 7, 2023. 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 some non-GAAP financial metrics during today's call. A detailed reconciliation of GAAP and non-GAAP metrics can be found under the financial portion of the Investor Relations section of akamai.com. And with that, I'll turn the call over to Tom.

Thanks, Tom, and thank you all for joining us today. I'm pleased to report that Akamai delivered excellent results in the third quarter with revenue, operating margin, and earnings all exceeding the high end of our guidance range. Revenue grew to $965 million in Q3, up 9% year-over-year. Non-GAAP operating margin was 31%, and non-GAAP earnings per share was $1.63, up 29% year-over-year. Ed will cover the key factors that drove our bottom line performance in his portion of the call. I'll now say a few words about each of our three main product areas, starting with security, our largest source of revenue. Security revenue grew 20% year-over-year in Q3. The acceleration in security growth was driven in part by especially strong demand for our market-leading Guardicor segmentation solution as enterprises confront the ever-present threats from malware and especially ransomware. CISOs and corporate boards everywhere have seen the recent headlines about devastating ransomware attacks, including at two casino hotels in Las Vegas, a major manufacturer of cleaning products in the U.S., and at a multinational provider of systems for smart buildings. One of them reportedly paid $15 million to get ransomware out of their systems. Another is reportedly spending $25 million to deal with the aftereffects, and a third has reported more than $100 million in losses from the attack. Customers who purchased our segmentation solution last quarter include a major global services provider, one of the world's most recognized entertainment brands, and a leading bank in Switzerland that renewed their segmentation protection with a significant upgrade. We also saw strong demand for our market-leading web app firewall solutions in Q3, where we continue to win against competitors who are challenged to provide the levels of reliability and performance required by major enterprises. Customers who switched to Akamai, including a nationwide retail chain in the U.S. and a leading global manufacturer based in India, also told us that our competitors simply can't provide the level of support and professional services that they need and have come to depend on from Akamai. Customers also value being able to purchase an entire suite of integrated security products from Akamai, a provider who they trust to keep them safe against a wide variety of attacks. For example, we're seeing very strong interest in our new API security solution that we announced last quarter. This new product is in its very early days, but we've already integrated it with our market-leading web app firewall solution to make it even easier for customers to implement. At the Black Hat security conference last quarter, Akamai API Security was named one of the 20 hottest new cybersecurity tools by CRN, a major trade publication for channel resellers. And as with Guardicore, you don't need to be a CDN customer to benefit from this new solution. Turning now to cloud computing. I'm pleased to say that we're on track with our product development, infrastructure deployment, and conversations with customers about use cases well suited for the Akamai connected cloud. Since our last call, we've gone live with seven more core compute regions in Amsterdam, Jakarta, Los Angeles, Miami, Milan, Osaka, and Sao Paulo. In addition to the six that we opened earlier this year and the 11 that we acquired from Linode, this brings our total to 24 core compute regions to serve Akamai connected cloud customers. Of course, there are other cloud companies with a few dozen data centers. But Akamai is unique in having these data centers interconnected to the world's most distributed edge platform with more than 4,100 points of presence across 750 cities and 130 countries. As one trade publication wrote recently, Akamai is focusing on a future where scale becomes more about the size of the network versus the size of its data centers, more effectively powering modern applications. We agree. Akamai's massively distributed edge network, 25 years in the making and managed by Akamai's team of experts around the world, is a key differentiator in our strategy. We believe that next-generation applications will need next-generation cloud infrastructure, and we intend to chart the course for the next decade of cloud computing when more of the compute will be done closer to the end user and where we believe our platform will have an important edge over more centralized models. As IDC put it in July, Akamai brings the simplicity, affordability, and accessibility of its cloud computing services to larger commercial customers on an architecture built for the next decade, not the last. The Akamai Connected Cloud will put containers and VMs closer to end users and bring enterprise workloads to locations around the world that are otherwise difficult for organizations to reach. Customers are responding to Akamai's unique offering, and we've already gained cloud computing business across multiple verticals in every major geography, including a European streaming media company, a digital advertising company in Japan, a large financial institution in Indonesia, an e-commerce platform in Korea, major carriers in EMEA and Central America, and a television network in South America. In addition to direct sales, we're seeing good traction in our cloud computing partner ecosystem, where we're acquiring new customers by selling with cloud service providers and managed service providers. We've also partnered successfully with independent software vendors and SaaS and PaaS providers. In fact, we recently signed one of the world's best-known SaaS providers and our second largest cloud computing deal since we acquired Linode. Turning now to content delivery. I'm pleased to report that we saw an acceleration of traffic growth in Q3. In addition, we acquired enterprise customer contracts from StackPath and Lumen Technologies following their decisions to exit the CDN market. As Ed will talk about shortly, the financial terms of the acquisitions were very attractive for Akamai shareholders. In summary, we are very pleased by our performance in Q3. Our expanded security portfolio is deepening our relationships with customers, our cloud computing plans are executing on schedule, and we continue to invest in Akamai's future growth while also enhancing our profitability. Now I'll turn the call over to Ed for more on our Q3 results and our outlook for Q4 and the full year. Ed?

Thank you, Tom. As Tom mentioned, Akamai delivered a strong and very profitable quarter in Q3. In my remarks today, I'll cover our Q3 results and then provide some perspective on Q4, share some details on our recent customer contract acquisitions, and close with our increased full year 2023 guidance. First, let's discuss revenue. Total revenue for the third quarter was $965 million, up 9% year-over-year as reported and in constant currency. In the third quarter, Security revenue was $456 million, growing 20% year-over-year as reported and 19% in constant currency. Security revenue growth was primarily driven by continued strength in our segmentation product, which is now over $100 million on an annualized run rate basis and up 97% year-over-year. I'll note that during the quarter, we had approximately $6 million of one-time segmentation license revenue. Adjusting for that one-time license revenue, total security growth for the third quarter would have been 18% year-over-year as reported and 17% in constant currency. And segmentation revenue growth would have been approximately 62% year-over-year and 60% in constant currency. In addition to strength in segmentation, we also saw very strong growth in our flagship Web Application Firewall or WAF product family. This growth was primarily driven by stronger-than-expected adoption of new security bundles offered to new and existing customers that we introduced this year. The new security bundles include additional security entitlements such as more security policies, additional security configurations, and more advanced rate control policies. Many existing customers are seeing greater value in these new bundles and as a result, are spending more with us. Moving to compute. Revenue was $130 million, growing 19% year-over-year as reported and in constant currency. On a combined basis, our securities and compute business product lines represented 61% of total revenue, growing 20% year-over-year and 19% in constant currency. Shifting to delivery. Revenue was $379 million, declining 4% year-over-year as reported and in constant currency. It's worth noting that delivery was aided by approximately $4 million in revenue from the selected CDN customer contracts we acquired from StackPath. International revenue was $467 million, up 11% year-over-year and up 9% in constant currency. Foreign exchange fluctuations had a negative impact on revenue of $3 million on a sequential basis and a positive $7 million benefit on a year-over-year basis. Moving now to company profitability. Non-GAAP net income was $251 million or $1.63 of earnings per diluted share, up 29% year-over-year and up 28% in constant currency. These especially strong EPS results exceeded the high end of our guidance range by $0.11 and were driven primarily by higher revenues and continued progress on the cost savings initiatives we outlined over the last few quarters. As an example, we continue to reduce our third-party cloud spend by migrating internal workloads to our connected cloud platform. In Q3, our third-party cloud spend declined 26% year-over-year. Moving to margins. Our cash gross margin was 73%. Included in our Q3 cost of goods sold was approximately $5 million of transition services agreement or TSA costs paid to StackPath. With customer contract acquisitions, TSA payments are used to cover the seller's customer-related network and support costs during the migration period. I'll provide further detail on expected TSA costs going forward in the guidance section in a few moments. Adjusted EBITDA margin was 43%, and our non-GAAP operating margin was 31%, two points ahead of our guidance, driven by our revenue outperformance and continued focus on driving down costs across the business. Moving now to cash and our use of capital. As of September 30, our cash, cash equivalents, and marketable securities totaled approximately $2.1 billion, which includes the proceeds from the convertible debt raise we did during the quarter. As a reminder, in August, we issued $1.65 billion of senior unsecured convertible debt that will mature on February 15, 2029. The notes will bear interest at a rate of 1.125% per year payable semiannually. Finally, the net proceeds of approximately $1 billion from this offering have been invested in highly liquid marketable securities. These securities yield approximately 5.25% on a weighted average basis with maturities close to May 2025 as we intend to use these proceeds to pay off approximately $1.15 billion of convertible notes that mature in May 2025. For the third quarter, we spent roughly $113 million to repurchase approximately 1.1 million shares. We now have roughly $600 million remaining on our previously announced share buyback authorization. Our approach to capital allocation remains the same to opportunistically buy back shares to offset dilution from employee equity programs over time while maintaining sufficient capital to deploy when strategic M&A presents itself. Before I cover Q4 guidance, I want to provide a quick reminder about our typical fourth quarter dynamics and add some color to our two recent transactions with StackPath and Lumen. As in prior year, seasonality plays a significant role in determining our financial performance for the fourth quarter. Typically, we see higher-than-normal traffic from large media customers and they pick up in seasonal online retail activity from our e-commerce customers. Both of these traffic patterns are difficult to predict. Q4 also tends to have higher operating expenses than in Q3, driven by higher sales commissions due to accelerator payments for sales reps who overachieved their annual quotas. As it relates to the transactions with StackPath and Lumen, first, both transactions were acquisitions of selected CDN customer contracts, including over 200 net new customers to Akamai. We did not acquire any other assets or liabilities at either company. Second, we expect the two transactions combined will add approximately $17 million to $20 million of revenue in Q4. Third, we expect to record approximately $13 million to $14 million of StackPath and Lumen TSA costs in Q4. These costs will be recorded in our cost of goods sold and will have a negative impact of approximately 1 percentage point on gross margin, adjusted EBITDA margin, and non-GAAP operating margin. Combined, the StackPath and Lumen TSAs will negatively impact our Q4 EPS by approximately $0.06 to $0.07. We do not expect to incur any material TSA costs in 2024. And finally, our expectations for these customer acquisitions remain the same as we disclosed previously for the full year 2024. As a reminder, we expect the customer contracts acquired from StackPath to add approximately $20 million of revenue in 2024 and to be accretive to non-GAAP earnings per share by $0.03 to $0.05. And we expect the customer contracts acquired from Lumen to add approximately $40 million to $50 million of revenue in 2024 and to be $0.08 to $0.12 accretive to non-GAAP EPS. With all that in mind, we are now projecting fourth quarter revenue in the range of $985 million to $1.005 billion, or up 6% to 8% as reported and in constant currency over Q4 2022. At current spot rates, foreign exchange fluctuations are expected to have a negative $8 million impact on Q4 revenue compared to Q3 levels and a positive $2 million impact year-over-year. Taking into account the impact of the StackPath and Lumen TSAs for the fourth quarter, we expect cash gross margins of approximately 72%. Q4 non-GAAP operating expenses are projected to be $305 million to $311 million. We expect Q4 adjusted EBITDA margin of approximately 41%. We expect non-GAAP depreciation expense to be between $123 million to $125 million, and we expect a non-GAAP operating margin of approximately 29% for Q4. Moving on to CapEx. We expect to spend approximately $143 million to $153 million, excluding equity compensation and capitalized interest in the fourth quarter. This represents approximately 15% of our projected total revenue for the fourth quarter. Additionally, our CapEx guidance includes the integration requirements to support the traffic for both CDN customer contract acquisitions. Based on our expectations for revenue and costs, we expect Q4 non-GAAP EPS to be $1.57 to $1.62. This EPS guidance assumes taxes of $50 million to $52 million based on an estimated quarterly non-GAAP tax rate of approximately 17%. It also reflects a fully diluted share count of approximately 155 million shares. Looking ahead to the full year, we have increased revenue to a range of $3.802 billion to $3.822 billion, which is up 5% to 6% year-over-year as reported and 6% in constant currency. At current spot rates, our guidance assumes foreign exchange will have a negative $18 million impact on revenue in 2023 on a year-over-year basis. We are raising our security revenue growth expectations to approximately 15% for the full year 2023, and we continue to expect to achieve approximately $0.5 billion in revenue from compute in 2023. And despite a year of significant investment, we are estimating non-GAAP operating margin of approximately 29%. With all that in mind, we have raised our estimated non-GAAP earnings per diluted share to a range of $6.08 to $6.13. Our non-GAAP earnings guidance is based on a non-GAAP effective tax rate of approximately 17%, and a fully diluted share count of approximately 155 million shares. Finally, our full year CapEx is expected to be 19% of total revenue. In closing, we are very pleased with how the business is performing in 2023 as we continue to invest for revenue growth and improve our profitability. With that, we now look forward to your questions.

Really nice quarter there. Understanding you're selling Guardicore to the installed base primarily. But what are you seeing with selling outside of the installed base with security and compute and specifically within security? Is there a way to think about that drag along effect or how the Guardicore enterprise sales team are really dragging along the rest of the security or even compute portfolios into any net new customer wins? Really, the crux of the question is how is the sales process going outside of the CDN installed base?

Jim, this is Ed. I'll start. Tom, you can jump in if there's anything you want to add. So actually, it's interesting, you started off by saying Guardicore is mostly to the installed base. Actually, we've done a really nice job of selling to new customers. One of the things that came over as part of the Guardicore acquisition is a pretty robust channel, and pretty much every deal is done through the channel. As a matter of fact, I think we can do a better job of selling in the installed base and probably come up with some new incentives to incentivize the sales force to do that. So there's a ton of room in the installed base because most of the growth in Guardicore has come from outside the installed base.

Yes. And I would just add to that, that there is good drag along with Guardicore, and for example, Enterprise Application Access is considered north-south, the other east-west; both are really important in terms of keeping malware out and identifying when it gets in and really blocking the spread. So it's good that way, too.

Helpful, guys. And just a follow-up, Tom, you had actually started to allude to it a little bit. But as we think about those Lumen and StackPath contracts, I guess how much wallet share do you have in aggregate of these new customers? Or what's the overall opportunity for those specific customers beyond just that CDN revenue that you bought?

Yes. Well, obviously, there's a CDN revenue, which is what transfers, but we're going to be looking to sell our security and compute solutions into those 200-plus new customers. And I think there's some good opportunity there.

Good afternoon. Thank you for taking my question. Tom, maybe I'll start with you. I want to have a broader conversation regarding some of these contracts that you've been acquiring. It's setting up to be a little bit of a pattern. So look, you've been doing delivery for the better part of two decades here. I wanted to get your kind of longitudinal perspective on what you're seeing in the marketplace and the market backdrop that's pointing more and more towards consolidation in the delivery here. And then ultimately, I wanted to get your perspective on how you think this is going to impact some of the pricing dynamics and ultimately, your pricing power in the delivery space. And then I have a follow-up for Ed, if I may.

I wouldn't characterize it as a pattern. It's been quite some time since we've acquired a competitor or their contracts. In this case, two CDNs have chosen to cease operations while still remaining as companies. They wanted to ensure their customers have the best experience as they exit the CDN space, so they approached Akamai, recognizing our leadership in this area, and presented very attractive financial terms for the transaction. This decision benefits our shareholders by allowing us to take on these customers and also provides an opportunity to sell them security and compute services, which is advantageous for them. We're not actively seeking to acquire other CDNs, but compelling opportunities for shareholders do arise occasionally. I don't foresee any significant shifts in the marketplace; companies like hyperscalers and several others continue to offer delivery services, so I doubt there will be any fundamental changes or pricing shifts. Additionally, as we've mentioned, Akamai is taking a more conservative approach to pricing. We're avoiding variable traffic that doesn't make financial sense because our focus has shifted towards investing in our compute offering, which has significant growth potential.

Perfect. Thank you. Ed, I was hoping I could parse out some of your prepared remarks with respect to the third-party cloud spend that you're effectively in-sourcing. I believe a couple of quarters ago, you may have ballparked that figure at about $100 million. So please correct me if that recollection is correct. And also just wanted to get a sense of how far down the path you are in this in-sourcing of third-party cloud spend on your Connected Cloud platform. Thank you.

Yes, sure. That's a great recollection, Fatima; you're correct. That's about $100 million or actually north of $100 million. So we are still in the earlier innings of the journey. However, we've made a lot of progress, and the team is doing a great job. So we expect to see that the savings continue to ramp. I'm very happy to say that we're slightly ahead of where we expected to be, and we have a lot of confidence that we'll be able to drive the type of savings that we expect to throughout next year.

Thank you.

Thanks for taking my question. Just want to dive into security a bit and see are the trends differing between international and domestic in terms of what products are getting better versus not? And then one follow-up question after that. Thanks.

I don't think there's a fundamental difference. The attacks are global in nature and the same attacks we see here domestically. Yes, you do see a little bit more attacks where there are hotspots, wars taking place, or political tensions, there'll be more attacks. But the nature of the attacks is similar. You've got denial of service attacks, ransomware, application layer attacks, and more recently, AI attacks. So that happens everywhere because there's no reason it should be in one geography versus another.

Got it. And then for the Connected Cloud and just the compute segment as well, you guys talked about a lot of really nice international deals, but also just wanted to get a pulse on what's happening domestically and appetite for the products here.

Again, I think that is universal as well. We'll have special advantages in locations where the hyperscalers aren't. But that kind of compute capability can be accessed by customers anywhere. Big U.S. companies care a lot about being able to give really good performance for their users all around the world. So again, I think there's not a fundamental difference between the U.S. or other regions. We have gotten off to a really good start, and I would say APJ, but it's across the board. We're deep in conversations with major enterprises across all the major geographies.

Thanks for taking the questions. And Ed, I appreciate the color on the additional bundles of security and the additional services you added to those bundles. And then in our conversations, we did pick up a decent price up with when those services are added. So can you help us understand what sort of pricing uplift, if any, there might be there? And what the penetration rate is in your installed base currently of those services and what the opportunity is going forward?

Yes. I'll begin, and then Tom can add if he has anything to say. We have implemented a targeted program aimed at customers in specific industries such as commerce, manufacturing, health care, pharmaceuticals, and financial services, which we typically refer to as our legacy web customers. As we have discussed, we have incorporated many additional services and functionalities into our offerings. When it comes to renewals, we are able to upsell. There are likely around 2,000 to 3,000 customers in this targeted group, and we are approximately halfway through the renewal process. Keep in mind that our average customer contract length is about 18 months, so this initiative may span around 18 months to 2 years. We are quite pleased with the average increase in sale price that we are observing.

Great. And then maybe if I could, just on compute. Now that the majority of those core data centers are online, can you talk about how much of the contracted space you filled out and what sort of utilization you're hoping to achieve as you enter 2024? I know you're not going to provide any sort of guidance here, but any sort of guidepost in terms of the actual capacity that will be online in '24 and what the compute pipeline looks like heading into next year to fill that capacity would be helpful just to understand the capital needs of that business as we think about '24.

Yes. In terms of the capital, we pretty much have now done the major buildout. And of course, we're big consumers of that ourselves as we move our own applications in-house to Akamai Connected Cloud. And we've also got plenty of room to take on major enterprise business. So I don't think you'll see a lot of CapEx associated with the big core data centers until we start generating a lot more revenue from that. And then we would build out further. There will probably be another couple that we do. What you will see is a relatively smaller amount of CapEx as we do build out into our existing Edge pops. And our goal is to start equipping them with compute so that you can run containers, VMs, Kubernetes, and many more cities around the world. And more cities than you can do that with, for example, the hyperscalers. And so that will be taking place over the course of the next year, but it's not a large amount of CapEx, so much smaller than what you saw this year. And then going from there, it will depend on how fast the revenue grows. So it will be a good news story if we're back next year saying that, okay, great, we filled that up, and now we're going to be building out some more. So it's a much better situation than we were in this year where there was a lot of buildout getting ready and no revenue yet.

Great, thank you. Can you give us some color on sort of the nature of the compute deals that you're getting now versus maybe 6 to 12 months ago? And then what have you learned by putting some of your own enterprise-level workloads on the compute platform that you've maybe been able to utilize as you sell to customers? And how has that process benefited the product?

Yes, that's a great question. If we look back more than six months, there weren't many compute deals because the Linode infrastructure wasn't fully developed for that purpose. We had some initial trials, but now we're capable of handling mission-critical applications for large enterprises, and we're seeing positive growth in that area. We've gained a lot of insights from our experiences. On the positive side, we are saving significant costs, which will also benefit our customers. We're experiencing strong performance that often surpasses what we see from hyperscalers. Looking ahead, we will have improved scalability and faster responsiveness, which is crucial for Akamai due to the unpredictable nature of peak events and large customer demands. Our platform is well-suited for handling these scenarios with security and delivery. We're now extending those capabilities to compute, allowing us to quickly deploy additional compute instances. Overall, we're very satisfied with our own cloud usage. However, we also learned that transitioning from a hyperscaler to Akamai isn't simply a matter of flipping a switch. There is a process involved, much like any other cloud migration, but the effort is very worthwhile. We're becoming excellent references for potential customers, and we're collaborating with various cloud partners to assist their clients in migrating to Akamai Connected Cloud.

Could you provide an update on the backlog and customer interest? It seems to be progressing well. Also, where do we stand with the value-added services and tools, and the overall cloud portfolio? Lastly, could you outline what the long-term margins of this business are expected to be? We often receive questions about how you can offer this at a lower price than the cloud providers and whether that will affect your margins.

Okay. A lot of components there. I'll start with some of them, and then probably Ed will fill in with some. Yes, we're in a lot of really good conversations, as you can imagine, because the world's major media companies, gaming companies, and commerce companies all use Akamai. They have for many, many years. They trust us for scale, reliability, and performance; they trust us with security. They trust us not to compete with them. As they see what we're building in terms of compute, they like the idea of getting better performance, more distributed compute capabilities. And some of those folks really like the idea of a much lower price point. And as an extra benefit, especially if you're in media or commerce, it's getting to be a bigger problem that they're cutting such giant checks to their leading competitor and sharing all the crown jewels of their data with their leading competitor. So they're taking the Akamai solution with great interest, I would say. Now in terms of being cheaper, Akamai has been for a long time the world's most distributed platform. We run one of the world's largest backbones. We have worked for many, many years to be incredibly efficient in terms of moving data around and doing the delivery. And that gives us a real advantage of being able to do this now for compute in a very cost-effective way. Now that said, I am sure that the hyperscalers pay a little bit less for their hardware than we do, probably not a lot less, maybe a little less. But when it comes to everything else, I think Akamai is in an excellent position. What we're seeing in the marketplace is that they'll get their best offer from the hyperscalers, and we can be a lot lower than that and be very profitable at doing it. The good news, I think, for Akamai is that here you've got a $10-plus billion market growing at 20% a year. And we're a tiny guy there compared to the hyperscalers. We can operate at a level that is not threatening to them in any way. They have to worry about each other. There's plenty of room for us to take on a lot of revenue at lower price points and very good margins for Akamai. Now in terms of the marketplace, that's an area, obviously, with the hyperscalers are way ahead some of those folks, every application ever made is available in the marketplace as a managed service. We are growing our marketplace. We are growing the tools that are available on Akamai Connected Cloud. The first applications that we're targeting are applications that are more easily able to be cloud agnostic, that are not locked in, that aren't using 20 other applications in the marketplace. We don't need all those applications. All we need is to get going as a tiny share of that market. We've identified just, for example, in the media vertical alone, there's a lot of applications that are amenable to moving to our platform. A lot of our partners in our marketplace to begin with are media-related companies, for that reason because they're also threatened by the hyperscalers, and they're very excited about having media applications beyond Akamai. Ed, do you want to add anything there?

Yes. Just a couple of things. As Tom talked a little bit about the leverage we have with the backbone, but we also have a lot of other leverage in the company with our go-to-market, where the focus is going to be initially with our installed base. As Tom talked about in media, to start, there's a tremendous amount. We pretty much work with every major brand. So there's a lot of leverage from that perspective. Also, the people that build and deploy the network are the same people who are building and deploying our CDN network. We're getting leverage with our co-location vendors and things like that. I've seen some pretty large proposals go out that get us margins that are pretty similar to the company margins in terms of gross margins that are somewhere between security and delivery and operating margins that potentially could be even greater as we get scale to the bottom line. So there's an enormous amount of margins. If you think of the math that we're doing with how much we're saving, the amount of capital that we're deploying, and the cost, we're going to be saving a tremendous amount of money on moving our own applications, and we'll be able to offer some of that to our customers as well.

I'm surprised we haven't mentioned AI during this conference call. Can you discuss the impact of AI on your ability to incorporate it into computing? Do you think it's feasible to implement it at the edge or through your CDN Edge? Also, do you see it as a potential risk, given that while inference may be distributed, much of the computing might remain centralized, which could discourage customers from migrating their applications to your platform? How should we consider the AI opportunity alongside the challenges customers may face in making that transition?

Great question. There are many components to this topic. At a high level, there are significant opportunities. However, I want to take a step back. Akamai has been utilizing AI and machine learning in our products for a long time. These technologies are useful for tasks like anomaly detection, bot detection, verifying if the right person is accessing a bank account with valid credentials, and identifying malware infections within enterprise applications. We have multiple applications of AI and machine learning. With Gen AI, while it offers some advantages, it also facilitates attacks. It's now easier to adapt malware into various forms, complicating detection efforts. Our teams have quickly developed very sophisticated bots using Gen AI, and we're already observing increased penetration due to this technology. This indicates greater cybersecurity risks for enterprises; they are likely to face more breaches. It’s essential to enhance your defense strategies. Products like segmentation become even more crucial, as preventing the spread of breaches relies on quick detection. Our growth in this area is strong, and we have a market-leading solution. Regarding the impact of Gen AI on compute, I believe it will require more processing power over time, which is beneficial for vendors like Akamai that sell compute. There is a distinction between model generation, which involves large models and requires significant resources in core data centers, and inference engines, which can operate at the edge and are ideal for many applications. We have several partners already migrating their AI models to Akamai for inference purposes, and I anticipate they will be utilizing this in our marketplace. In fact, we are currently leveraging AI as we migrate our internal applications to Akamai Connected Cloud. I expect this will create substantial revenue over time due to various AI-related applications. As for risks, model generation will take place in large data centers, which is not a concern for Akamai since we have several of these facilities. That work will not happen at the edge. However, inference engines will largely function at the edge, and we have a strong advantage there, as few companies have edge capabilities comparable to Akamai.

This is Dua speaking for Abdul. I just wanted to ask for you broadly on the enterprise spend environment. I know we've previously noted elongating sales cycles. I was just generally curious whether there's any change there. If you had to characterize it, is enterprise IT spend incrementally worse, better, or about the same versus, let's say, 90 days ago?

Yes. Good question. I would say, I think there are a lot of companies that are being cautious. We have seen a slight uptick in bankruptcies as you typically would see in cycles like this. But we're not seeing a significant impact, certainly in our security business. If anything, we've seen better-than-expected results there. I think security is not as impacted, at least at the moment. Obviously, there's a lot of speculation out there that we're heading towards a recession and things can change pretty quickly. But so far, we fared very well. Also, if you think about our messaging around compute, one of the biggest challenges a lot of companies have is the runaway cost of compute, and we offer a very compelling option for people to look to save money and increase their performance by moving to us. So I think that will play into our favor in an environment like this.

Ed, how noticeable or how sudden is the increase that you're seeing in the sophistication of all these malware and ransomware attacks? The part of why I was asking is we noticed that you're launching some scrubbing centers in Canada. I'm wondering if that's driving CapEx a little higher to help make sure that you're able to address all these attacks. And then I have a quick follow-up.

Yes, let me just start on the product side, and then Ed will pick up your questions. So ransomware isn't related to scrubbing centers. Scrubbing centers are just restricting the flow of packets and screening out or scrubbing out the packets that are trying to flood any particular resource. That's nothing really per se to do with malware and ransomware; to filter that out, you need application layer defenses, where the scrubbing centers are routing layer defenses. Putting the scrubbing centers in Canada, and we're actually putting scrubbing centers in many more cities around the world and greatly increasing our capacity so that with local customers there, we can do the scrubbing for them locally. That gives them better performance while we're giving them the defense against the volumetric attacks. For ransomware, you need Guardicore; for malware, you need app and API security. Those are different products where they're done at our edge network in the 4,000 POPs, EdgePops we have around the world. And then, Ed, do you want to pick up the other part of that?

Yes, usually the establishment of the scrubbing center typically follows significant demand from customers. One reason for the slight increase in security this year is the rise in volumetric attacks in the healthcare sector, and to some extent, in the financial sector as well. This is fairly standard for us. Regarding capital expenditures or future builds, this is just routine; there's nothing particularly unusual to highlight, but we have noticed a slight increase in DDoS attacks, which are often episodic and tend to coincide with major high-profile incidents. Consequently, we typically see a surge in business, which often necessitates the construction of a scrubbing center. However, we are well aware of where the demand is increasing.

Okay, understood. And then, Ed, just as a follow-up, did you mention what was the total consideration paid for the acquired contracts from StackPath in Lumen? I'm wondering if those are expected to contribute something like $60 million to $70 million next year in aggregate. Are you assuming a similar ongoing run rate that those contracts had with the prior providers and extrapolating that into next year? Or are you contemplating into that any kind of expansion, contraction, or pricing that up or pricing that lower?

Sure. Let me break it down into different parts. Regarding any potential upsell with the more than 200 customers, I haven't included that in my projections. It would be additional upside if we can achieve it. Keep in mind that we selected specific contracts to purchase and chose not to take others, such as adult content and some contracts with small and medium businesses. Those are not included in our figures. If you have heard other numbers pertaining to these private companies, I would advise caution regarding their accuracy. We assessed the contracts we are acquiring and estimated how many will onboard, anticipated pricing, and the traffic retention, noting that some customers may switch, which could decrease our overall traffic. We aimed to account for all these variables. We're focusing on the integration of the contracts we acquired and determining the running rate for that business, considering the best estimates for volume and pricing dynamics. In our upcoming 10-Q, which will be filed tomorrow, you'll see that for StackPath, the initial contract fee is around $35 million, and there is a minor earnout. The TSA agreement also involves some complex accounting, as fair value must be assessed, and any excess costs will affect the purchase price. Consequently, the final numbers in the 10-K may appear slightly higher. For Lumen, the figure is approximately $75 million, with no earnout, but similar fair value considerations apply. Overall, that summarizes the agreements.

One wanted to follow up on two earlier questions that were asked, one on AI and then one on the contracts. Thinking specifically around AI, Tom, I appreciate your answer earlier. If we think about the opportunity to do inferencing at the edge, especially for cases like connecting devices or med tech or financial services now that people are increasingly worried about data and data residency. Can you speak to a little bit of your opportunity for that? And maybe alongside that, what investments do you need to make both in the software stack as well as in hardware infrastructure, be it GPUs or anything else to really capitalize and get your fair share of that opportunity? And then I've got a quick follow-up.

The issues of data residency and data sovereignty are increasingly significant, and Akamai is well-positioned to leverage this opportunity since we operate in 130 countries with our infrastructure. By moving compute capabilities to our edge locations, we can process data locally, which presents an exciting prospect for us. In the near future, we will be present in areas where even larger cloud providers do not have a presence. Currently, we are in the process of enhancing our software stack, and we have already entered the beta phase with several customers, indicating significant progress. Next year, we anticipate modest capital expenditures as we expand numerous edge locations to support compute capabilities. Although we currently support GPUs at our core data centers, for inference engines, we are successfully operating on CPUs. At the edge, where inference engines will be deployed, we don’t expect to use CPUs, as this will be more cost-effective compared to acquiring GPUs or specialized hardware. Such hardware may be more relevant for large-scale model development at core data centers, whereas edge deployment will be best served with CPUs for effective inferencing.

Got it. That's really helpful. And then just in terms of the StackPath and Lumen contracts, again, I appreciate the color in terms of your set of assumptions. Can you walk us through what you can do on your part to ensure those customers, whether they're net new or existing Akamai customers that maybe we're trying to use a multi-CDN approach, what you can do to get those customers to stay and to not turn off on to other competitors or even just kind of figure out how to reduce the peaty?

Yes, that's a great question. To start, we received an inbound request from these two companies, some of which are our long-standing customers. The first step in our relationship is to provide an orderly transition. This allows us to build a warm relationship with new customers and reinforces our connection with existing ones. These customers are already familiar with us, so instead of competing in an open market, we can meet with them to better understand their needs, contract details, pricing, and more. This sets the stage for a typical sales cycle. We have a strong track record of knowing our customers well. When considering the revenue distribution, we’ve had discussions with many of our customers and have established long-term relationships with many of them. Therefore, we are quite confident in the numbers we've provided. While changes can occur, our experience with both existing and new customers gives us confidence, especially since we have many additional services to offer that they haven't used before. With security being a major concern lately, many customers have expressed their satisfaction with our partnership.

Ed, I just want to quantify the TSA impact. It seems to be about 1.5 points impact cash gross margins in Q4. Is that accurate? And just to be clear, it sounds like that TSA ends at year-end? Or when exactly does that TSA add?

Yes. So there might be just a tiny bit that goes into Q1 just depending on how the migration goes. Hopefully, we can be done with it by the next couple of months. Yes, it's just under 1.5 points. I rounded down to a point, but you're right. If you just take the amount divided by the revenue, it's about 1.3%, 1.4%.

I understand. Looking at your delivery guidance, which seems to reflect your compute security revenue guidance, and factoring in the anticipated $17 million to $20 million revenue from Lumen and StackPath in the fourth quarter, it suggests that delivery revenue will be flat or possibly decline in Q4 compared to Q3. Is there any additional caution reflected in the Q4 delivery guidance due to what you're observing in traffic trends or any pricing pressure? Why aren’t we seeing the usual seasonal increase in delivery revenue, excluding those acquisitions?

Yes. That's a good question. First of all, there's $4 million of delivery in the Q4 number, so you have to look at the net delta between the two. Also, there's about an $8 million headwind from FX. There is some implied growth in the delivery number. As I talked about, there's a high degree of uncertainty in Q4 as it relates to traffic, right? Typically, we see a big seasonality in terms of retail and the media side of the business. Retail, as we've gone to zero overage, is less impactful. We still see some bursting. I think we're being probably a bit more cautious, especially on the commerce side of the equation for now. But there is some implied growth in there as you factor in those other two items I just mentioned.

Okay. That's helpful.

This is Yan Simoes on for Will. So first of all, just looking at delivery, how are you thinking about Q4 delivery trends this year versus normal seasonality given all the questions around the consumer and media activity in general? Any color there would be great.

Yes. As I just mentioned on the last question that we're probably a bit more cautious in terms of our outlook, certainly with retail. I mentioned earlier in an earlier question that we've seen an uptick in bankruptcies. We do have a lot of customers that are on the zero overage for commerce. So we're going into the season. We haven't seen it yet. It usually starts right after Thanksgiving. We're just being a bit more cautious. In terms of the media cycle, we did see a little bit of gaming activity. Gaming has been light for the last 1.5 years. A couple of years ago, we saw a big console refresh cycle. We're not expected to see that. So I'd say we're going into this quarter probably a bit more cautious than we normally have in Q4 and certainly what we've seen in the past.

First, earlier you talked about portability, particularly for cloud-agnostic workloads. I guess my question for you is, as you think about moving and garnering more share here, what are the steps that you could do to pretty much increase the ease of portability of workloads? As part of that, through your data center deployments, do you have essentially direct cross connect into the cloud on ramps of the major cloud providers to help facilitate the movement of workloads to your platform? And then I have a quick follow-up.

Yes. Great question. I think partners are really helpful there because they do a lot of the work in the first place to get into the third-party cloud provider. Today, some of them move between third-party cloud providers. There is a natural partner ecosystem that can be helpful to port that to Akamai. I think they're finding that the terms are even more favorable for them as well as for the partners. Also, as we grow the ecosystem of third-party capabilities, which with our qualified compute partner program, we're doing early focus on media there, which is our initial focus in terms of applications to move to Akamai. Yes, what it depends on the data center and the third-party cloud provider, but we do have direct connections in many cases. Of course, we operate, as I mentioned, one of the world's largest backbones and in a position to also make direct connections to major enterprises, which can help quite a bit. And then just as a quick follow-up. Just curious if you could talk a little bit about the M&A environment that you're seeing. I know you have that your security growth guidance that you gave at your Analyst Day, which includes some M&A. Just curious, any thoughts around the M&A environment, particularly for security? Maybe as part of that, just comments on valuation. Yes, valuations in the companies that we're most interested in are still extremely high. You see some of the recent acquisitions that have been done at very high revenue multiples. So not a lot of change yet. Probably at the fringes, it's getting harder. But for companies that we have an interest in, I'd not say it's made a lot of improvement yet. That may change with time. We just have to see. It's something we keep a close eye on. Of course, we're always looking, but we're very disciplined buyers. So it really has to make very good sense for our customers and for our shareholders.

Just one quick one on compute, kind of getting to the inflection or the acceleration in growth. I have a couple of questions. Just on sales cycles there, refractory, what is the typical sales cycle for a compute deal? And then secondly, with the bulk of the build-out done, and it sounds like functionality isn't the limiter, then it's really just getting to maturity of these sales cycles. The sales force has trained the functionalities there, the infrastructure is there. So it looks like Q4 builds in organic growth roughly similar to Q3. Trying to get a sense of timing of acceleration there.

Yes, that's a good question. There is indeed a sales cycle involved. Once a deal is closed, there’s the effort required to actually implement the service and the time needed for it to expand. It's not as straightforward as a typical service where things can be set up quickly; in this case, it might take a few months to fully migrate the app. The transition of traffic or compute resources won't happen all at once, and growth will be gradual. This is an ongoing process that you can expect to see develop throughout next year as we acquire new customers and increase their revenue. What we’ve noticed from our early signings this year is that they start with relatively small revenue, which then tends to grow over time.