Akamai Technologies, Inc. (AKAM) Earnings Call Transcript & Summary
February 20, 2025
Earnings Call Speaker Segments
Mark Stoutenberg
executiveHello, everyone, and welcome to Akamai's call covering results for the fourth quarter of 2024. Thank you for being here this afternoon. Joining us today are Tom Leighton, who serves as our Chief Executive Officer, and Ed McGowan, our Chief Financial Officer. I want to point out that what we share today contains forward-looking statements — this covers revenue and earnings guidance, our business outlook, goals spanning 3 to 5 years, and longer-term targets. Such forward-looking statements rest on present expectations and assumptions that carry specific risks and uncertainties, and they involve a variety of factors that might lead actual results to diverge materially from what those statements express or imply. Among these factors are, though not limited to, effects stemming from macroeconomic trends, the integration of any acquisition, geopolitical developments, and any additional risk factors laid out in our SEC filings. The forward-looking statements made on this call reflect the company's views as of February 20, 2025. Akamai assumes no duty to revise any forward-looking statements, which are valid only as of the date they are given. As a reminder, we will reference certain non-GAAP financial metrics throughout today's call. You can find a detailed reconciliation of GAAP to non-GAAP metrics under the financial portion of the Investor Relations section at akamai.com. Additionally, as part of our continued dedication to transparency, we've upgraded our disclosures so investors gain a more thorough grasp of our business. This quarter, we've put together a new presentation, accessible in the Investor Relations section of our website, that provides somewhat more detail on our product portfolios, financials, and performance goals. That presentation complements what's in our earnings release and annual filings, and we urge you to take a look at it. Inside the presentation, there's an overview of select revenue along with year-over-year revenue and growth rates. Keep in mind, every growth rate percentage is reported on a constant currency basis. With that said, I'll turn things over now to our CEO, Dr. Tom Leighton.
F. Leighton
executiveThanks, Mark. Today's press release shows that Akamai turned in a strong fourth quarter, with revenue of $1.02 billion and non-GAAP EPS of $1.66, which came in comfortably above the top end of the guidance we had provided. I'm happy to share as well that we've made great strides in our multiyear effort to remake Akamai from a CDN pioneer into a cybersecurity and cloud computing company that powers and protects business online. In 2024, for the first time ever at Akamai, Security generated the majority of our annual revenue, crossing the $2 billion mark and rising 16% year-over-year. Revenue from our cloud computing portfolio reached $630 million last year, a 25% increase over 2023. Of that figure, $230 million came from our cloud infrastructure services, which grew 32% over 2023. Our cloud infrastructure services are made up mainly of the compute and storage offerings we built on Linode. They also encompass our Edge Workers product along with ISV solutions that run on our cloud platform. Taken together, security and compute represented 2/3 of Akamai revenue in 2024, up 18% year-over-year. We also beat every one of our year-end annualized revenue run rate, or ARR, targets for the business's fastest-growing segments — specifically our Guardicore platform, our API Security solution, and enterprise revenue from our cloud infrastructure services. Those are 3 of the key areas we expect to fuel revenue acceleration across the overall business in 2026 and beyond; on the security side, Akamai has moved into new adjacent markets, expanding past point solutions to deliver a security offering that is more holistic and comprehensive. That has let us broaden our customer base and serve enterprises better with a wider portfolio that protects infrastructure, applications, APIs and user interactions across both cloud and on-prem environments. Q4 security growth came from sustained strong demand for Guardicore, our market-leading segmentation solution, as more enterprises turned to Akamai for defense against malware and ransomware. The Guardicore platform closed the year at an ARR of $190 million, up 31% year-over-year and ahead of our $180 million goal. During Q4, we signed our biggest Guardicore deployment yet with a leading IT services company in India. That solution spans 30,000 servers and nearly 300,000 endpoints. We also replaced a rival segmentation offering that was underperforming at major banks, in Hong Kong and in the U.S. Over 80% of our segmentation revenue in 2024 arrived via channel partners, among them one of the world's leading SIs, Deloitte, which builds its services and implementation expertise around our segmentation and API Security products to create value for customers that Deloitte knows well. In Q4, working together, we landed a $5.8 million contract with Petrobras in Brazil to lower the risk of a breach and ransomware attacks. We also teamed with Deloitte to help protect 2 large European banks from API risks. During the fourth quarter, Akamai closed a sizable API Security deal with a major U.S. asset management and brokerage organization. Finishing 2024, that offering reached $57 million in ARR — compared with merely $1 million when 2023 ended — and it surpassed the $50 million target we had set. Combined, our API Security offering plus the Guardicore platform wrapped up 2024 at $247 million in ARR. Looking forward, we anticipate these products will keep delivering robust expansion, and our objective is a 30% to 35% lift in their joint ARR across the year. We further expect that, over the coming several years, the swift expansion and increasingly material revenue contribution from these newer offerings will counteract slower growth in our more broadly adopted, market-leading Web Application Firewall and DDoS mitigation products. Consequently, and as described in the supplemental disclosure Mark referenced a few minutes back, we think we can sustain a roughly 10% CAGR in constant currency for our security offerings across the next 3 to 5 years, assuming typical M&A. That trajectory would take us past $3 billion in security revenue before the decade ends. Though our security lineup is delivering excellent results, our compute lineup is expanding at an even quicker rate and comes with a far bigger addressable market. We've traveled a great distance since entering the cloud computing space in 2022 by acquiring Linode. Last year brought substantial headway as we hit what we had targeted on revenue growth, new enterprise customer signings, infrastructure build-out, product development, partner ecosystem growth, and shifting our own applications off hyperscalers onto Akamai Cloud. Through Q4, compute customer signings kept up their fast tempo, encompassing 2 U.S. financial software firms, 2 of the biggest U.S. retailers, a European cybersecurity provider, an Asian enterprise software company, one of Southeast Asia's largest banks, and a Latin American intelligent transportation system provider. To gauge how our cloud infrastructure services are progressing, we've examined results via 2 lenses. The first lens, our main one, is adoption among bigger enterprise accounts — those spending beyond $100,000 in ARR. The second lens captures how these products perform across customers regardless of size. At year-end, roughly 300 enterprises were spending $100,000 or more in ARR on our cloud infrastructure services, a significant jump over the prior year. Together, those accounts closed the year at $115 million in ARR for our cloud infrastructure services, well above our $100 million objective. Counting every customer, ARR for our cloud infrastructure services ended 2024 at $259 million, a 35% increase year-over-year; that marks considerable progress since our 2022 Linode acquisition, when ARR from these services stood at about $127 million and was expanding in the mid-teens. Beyond helping us bring in more customers, the effort to turn Linode into something suited for enterprise use has let us shift several of our key offerings away from hyperscalers and onto Akamai's Cloud. The payoff has been better performance along with annual savings that comfortably exceed $100 million. A large number of our clients have likewise ramped up their consumption of our cloud infrastructure services considerably during the past year, with some increasing it by 4x or more. When the year closed, 15 customers were each paying above $1 million in ARR for those cloud infrastructure services — a figure that is more than triple what it was in 2023. Also, as of today, we've disclosed our first customer that has signed an agreement committing to spend in excess of $100 million on our cloud infrastructure services across the coming several years. In our view, this represents a striking confirmation of our new cloud capabilities, indicating that a highly sophisticated purchaser of cloud services trusts our capacity to deliver and to offer service and performance on par with, or superior to, the hyperscalers. The enhancements we've recently brought to our cloud platform will position us to serve even more enterprise customers in 2025; for instance, over the past year we grew Akamai's core data center footprint to 41 locations spanning 36 cities globally.
F. Leighton
executiveComing up next week, we intend to share news that our new managed container service is now live across our more than 4,300 points of presence, spanning over 700 cities globally. At present, we're running trials of this offering with customer workloads in excess of 100 cities; our object storage solution has received a major upgrade, delivering a 5x boost in scalability along with a 10x gain in performance, which puts it on par with the hyperscalers while carrying far lower egress fees thanks to the efficiencies built into our distinctive edge platform. GPUs were introduced by us to serve a range of AI and media use cases. A proof of concept was conducted by one customer comparing Akamai against a hyperscaler, after which they picked Akamai for text-to-image AI inferencing workloads. AI-powered speech recognition for its in-vehicle voice assistant is what another customer relies on our cloud for. And an OTT provider made the move from a hyperscaler over to Akamai, seeking a more cost-effective platform for its ad-supported streaming TV service; the scale and security of our Linode Kubernetes engine product were enhanced by us. Kubernetes platforms are operated by traditional cloud providers out of a comparatively small number of core data centers. The differentiated approach taken by Akamai will merge the computing power of our cloud platform with the proximity and efficiency of our edge, placing workloads nearer to users than any other cloud provider can. Leveraging technology acquired from Red Kubes in the prior year, we launched the Akamai app platform, which lets developers build and deploy highly distributed applications with just a few clicks; also, 9 compute ISV partners were added by us last year, lifting our total to 23. $36 million of cloud infrastructure services ARR at year-end was contributed by our ISV partners. Our opportunity for continued strong growth has us very excited, as we deliver the power of compute to the edge, with our broadly deployed network bringing compute instances closer to end users, via an open platform that ensures flexibility and portability, orchestrated resource deployment for efficient scaling and operations, and predictable pricing paired with an unmatched ability to minimize egress costs. An evaluation of public cloud platforms published last month by IDC, I think, captures our rapid progress in cloud computing nicely. Akamai was identified as a major player relative to industry peers in their worldwide public cloud marketscape for IaaS, with the observation that "Akamai has accelerated its journey into the public cloud IaaAS space, transforming from a pure-play CDN provider into a formidable public cloud competitor." Gartner also positioned Akamai as an emerging leader for Gen AI specialized infrastructure in their recent innovation guide for generative AI technologies. As we noted in the supplemental materials Mark mentioned, a growing number of AI use cases are being supported by us, with a special focus on inferencing. It's still early days, but the long-term revenue opportunity has us excited, and we believe that the unique properties of Akamai's Cloud position us to be a major player in AI inferencing in the years to come. Turning to the remainder of 2025, we're aiming to expand total cloud infrastructure services ARR at a rate of 40% to 45% on a constant currency basis. Our expectation is that enterprise customers will be the main engine behind the accelerating expansion of revenue from our cloud infrastructure services. Because we're enjoying such strong results in cloud infrastructure services, we intend to direct a larger share of our compute investments there. Specifically, we are currently shifting certain legacy cloud applications — those handling things like visitor prioritization, image and video management, and live streaming workflow — over to ISV partners who focus on those specialties, and we intend to move some or all of their workloads onto Akamai's Cloud. Beyond turning former rivals into valuable ISV partners for our cloud, we think this shift will let us devote more internal resources toward further building out and growing our cloud infrastructure services. This transition also means we anticipate that revenue from certain cloud applications will shrink during 2025. And, as Ed will explain shortly, we're forecasting roughly 15% growth across our cloud computing solutions overall in 2025. With cloud infrastructure services revenue continuing to climb quickly, we're confident we can speed up the overall cloud computing revenue growth rate again, reaching a CAGR of no less than 20% across the coming 3 to 5 years on a constant currency basis; that would turn cloud computing into our third $1 billion product line by 2027. Next I'll turn to content delivery, which remains a key profit generator that funds the development of new products driving future growth. Our distinctive edge platform — spanning more than 4,300 points of presence across over 700 cities — remains a major differentiator when it comes to reducing our costs, supporting enormous scale, and delivering superior performance; that holds true not just for delivery but for security and compute as well. On the security side, we leverage the platform to deliver a massive shield against every kind of attack without hurting performance or driving up costs. On the compute side, we leverage the platform to deliver function as a service through our Edge Workers product; and, as we'll reveal next week, we'll likewise leverage that same platform to operate our new managed container service across thousands of POPs in hundreds of cities; no one else in the market offers this, and it will let our customers place their compute workloads far nearer to users. Akamai realized significant cost synergies by relying on the same physical server to power our delivery, security, and now compute services across more than 4,300 POPs and 700 cities. That's a capability nobody else has, and a central reason Akamai has stayed so profitable while numerous competitors have faltered. The delivery customer base we have already installed keeps driving expansion in our security and cloud computing businesses too, since we reap the benefits of both competitive edge and performance by selling delivery, security, and compute together as one package on a single platform. That combined effect pays off particularly for security and compute clients who treat delivery as a feature they need and view it as essential to staying with us rather than going to other vendors; during Q4, we closed numerous agreements that bundled security and compute offerings together with our top-tier delivery. We also regained delivery business from rivals at a top AI technology company and at a major streaming media firm. Then in December, we bought chosen customer contracts from Edgio so that those customers could receive our industry-leading delivery services and have the chance to benefit from the entire Akamai portfolio of security and cloud solutions. I'm happy to say that early signals of recovery are appearing in the delivery market, with more clients ready to commit to multiyear agreements featuring predictable pricing, a generally more settled pricing landscape, and initial indications that traffic growth is leveling off. Consequently, and as Ed will explain shortly, we anticipate the year-over-year drop in delivery revenue narrowing to roughly 10% for this year. Should these positive trends persist, the decrease in delivery revenue ought to keep diminishing in 2026 and afterward. During our call last May, we pointed out that our biggest customer is working through political difficulties and is following a DIY approach.
F. Leighton
executiveSo because of that, we anticipate this customer's revenue will create roughly a 1% to 2% yearly drag on our total revenue growth rate across the coming couple of years, then level off at an amount comparable to certain hyperscaler clients of ours — approximately 2% to 3% of overall revenue. Even so, I'm happy to share that during Q4 we signed a 5-year committed arrangement with this customer, and it carries a meaningful minimum annual spend, giving us better predictability and lowering our exposure to their political circumstances in the U.S. Although we're glad about the strides made in the past year along our multiyear transformation path, there's still more for us to do when it comes to reaching additional new customers and selling our newer security and cloud computing capabilities into our existing client base. To generate stronger top line growth across the coming 3 to 5 years, we're reworking our go-to-market approach so that more resources line up with the faster-growing parts of our business, and so that we pick up the pace of adding new customers. Specifically, we've already started increasing the hunters-to-farmers ratio within sales, and adding more specialized sellers plus presales resources backing sales of our Guardicore platform, API Security and cloud infrastructure services. On top of that, we're putting more investment into partner enablement, since the channel has grown into a key driver of revenue growth for us. Following guidance from one of the world's top consulting firms, we're additionally launching a significant project aimed at optimizing our sales operating model, account coverage framework, compensation structure, pricing strategy and how we make use of our channel partner ecosystem. As disclosed in the supplemental forecast posted on Akamai's Investor Relations website, we think that combining double-digit security growth, extremely rapid growth in cloud computing, a delivery business that's stabilizing and a product mix that keeps improving ought to let us speed up revenue growth in the years to come and reach double-digit revenue growth by the end of the decade, or possibly earlier; actually, once you strip out the effect of foreign exchange headwinds and the large customer I referenced before, it's clear the acceleration has already begun. Setting aside those 2 factors, revenue growth sped up in 2024 compared with 2023; and as Ed will explain soon, we expect even more acceleration in 2025. We think that boosting our top line growth and product mix, together with our ongoing push to gain efficiency, will assist in lifting operating margins, allowing us to hit and then surpass our 30% target over the coming several years. We further think we can get back to growing our non-GAAP EPS in 2026. Though plenty of work remains, we're quite optimistic about what lies ahead. Our cloud computing strategy is gaining traction just as we pictured it. Our broadened security portfolio is letting us deepen and widen our ties with customers and partners. We're still putting money into Akamai's future expansion, and at the same time we're keeping profitability at a solid level. At this point I'll hand things over to Ed, who will share more about how we did and what we expect going forward. Ed?
Edward McGowan
executiveThanks, Tom. Before I get started, I'd like to echo what Mark said at the top of the call and draw attention to a few new disclosures that came out earlier today. The materials we put up in the IR section of our site contain somewhat more detail than we'll walk through in today's prepared remarks. What we're trying to do is give you a deeper look into the business and lay out our refreshed long-term objectives. We don't plan to give this much disclosure each quarter, but now and then we'll add extra context when we think it's useful. I'll start by covering what happened in Q4, then share some perspective on 2025 — including several new items we're disclosing — after that I'll take you through guidance for Q1 and the full year 2025, and wrap up with how we think about revenue and profitability targets over the long run. Turning now to Q4, let's begin with revenue. We posted total revenue of $1.020 billion, which represents 3% growth compared with a year earlier, both as reported and on a constant currency basis; throughout the fourth quarter, our compute and security portfolios continued to deliver healthy growth. Edgio contributed about $9 million in revenue during the quarter, right in line with our expectations. Compute revenue climbed to $167 million, marking a 24% rise year-over-year as reported and 25% on a constant currency basis. For the fourth quarter, our compute revenue split into $65 million from cloud infrastructure services and $102 million from our other cloud applications. One of our biggest customers, as Tom mentioned, has made a commitment to spend no less than $100 million on our cloud infrastructure services across the coming few years. Their workloads should begin ramping up by the end of 2025, in our expectation. Now to security revenue. We saw $535 million in security revenue, which grew 14% year-over-year both as reported and in constant currency. In Q4, we recorded roughly $12 million of onetime license revenue, versus $5 million during Q4 of last year. Like we mentioned in the additional disclosure we provided, our security revenue broke down into two parts: $205 million came from Zero Trust Enterprise along with API Security, and that figure rose 51% in constant currency compared to the same period last year. The remaining $1.84 billion came from all our other security offerings, and those climbed 14% in constant currency year-over-year. When you put compute and security together, revenue was up 16% year-over-year on a reported basis and 17% in constant currency for Q4, and that combined category now makes up 69% of our total revenue. On the delivery side, we brought in $318 million, which was a bit better than what we had expected, though it declined 18% year-over-year both as reported and in constant currency. Looking ahead to 2025, we expect delivery revenue to improve, and that's based on a more favorable view of delivery, which Tom talked about earlier. Overseas sales totaled $490 million, which is a 2% increase compared to the same period last year, or 4% when you strip out currency effects. That accounted for 48% of our overall revenue in Q4. And then on the currency side, sequential revenue took an $8 million hit from forex swings, while the year-over-year drag was $6 million. Now turning to profitability. Our non-GAAP net income for Q4 was $254 million, which works out to $1.66 per diluted share. That's down 2% versus last year, flat in constant currency, and it came in well above the high end of the guidance range we'd given. A few factors primarily drove these EPS figures past our guidance: revenue that came in a bit stronger than anticipated, transition services—or TSA—costs tied to the Edgio transaction that were lower than expected, bigger savings than forecast from the headcount reductions we told you about in Q3, and reduced payroll expenses because some hiring slid from Q4 into Q1. Lastly, capital expenditures for Q4 totaled $193 million, which works out to 19% of revenue. Now let's turn to how we're allocating capital. Over the fourth quarter, we used roughly $138 million to repurchase about 1.4 million shares. Across the entire year, we allocated $557 million toward buying back approximately 5.6 million shares. When 2024 wrapped up, roughly $2 billion remained accessible through the repurchase authorization we already had in place. One key thing to note: over the past ten years, we've hit our objective of buying back stock to offset dilution caused by employee equity programs, and on top of that, we brought our shares outstanding down by approximately 16% across that stretch. Going forward, nothing has shifted in how we think about capital allocation: we'll continue buying back shares as time goes on to neutralize dilution from employee equity programs, while remaining ready to act when opportunities arise in both M&A and share repurchases. Before turning to guidance, I want to highlight a couple of items that should assist you with your 2025 models. The first relates to revenue tied to the Edgio transaction. We now anticipate Edgio revenue of roughly $85 million to $105 million for 2025. Beyond that, we expect around $6 million in TSA costs during Q1 tied to Edgio. Roughly $4 million of that will sit in cost of goods sold, with the rest appearing in operating expense. We don't expect any TSA costs after Q1. Second, we anticipate CapEx will come in about 2 percentage points above last year's level for the reasons that follow. First, we intend to boost spending by roughly 1% of revenue to handle the additional traffic generated by the Edgio transaction. Additionally, roughly one more percent of our revenue should be allocated to building out infrastructure in specific regions tied to that $100 million cloud infrastructure services deal we shared earlier today. Beyond that, we anticipate the bulk of our CapEx outlays will land early in the year, and Q1 will carry much heavier spending compared with the quarters that follow. The reasons are twofold: the items I just referenced, plus our intention to shift $10 million to $15 million of CapEx forward into Q1, which helps us limit how much we're exposed to possible tariffs that might get announced later this year. Third, we expect interest income to decrease in 2025, driven by lower cash balances following recent acquisitions, our intention to retire the $1.15 billion convertible debt instrument maturing on May 1, 2025, and anticipated lower investment yields as interest rates decline over the course of the year. Fourth, when it comes to 2025, we expect that volatility will stay elevated in currency markets abroad, caused by shifts in Federal Reserve policy that are hard to predict in both timing and size, plus how those shifts affect interest rates. To show just how big an effect FX can have on what we report. From our most recent earnings call, which happened in early November, the U.S. dollar's strength has hurt what we expect for 2025 by roughly $18 million in revenue calculated on an annual basis, and it has hurt our non-GAAP operating margin by roughly 30 basis points, and it has hurt our non-GAAP EPS expectations by roughly $0.09 per share. One thing to keep in mind about currency: we bring in about $1.2 billion of annual revenue that comes from foreign currency, and the euro, yen and Great British pound are the biggest sources of revenue that aren't in U.S. dollars. On top of that, our costs that aren't in U.S. dollars tend to be much lower than revenue and are mostly in Indian rupee, Israeli shekel and Polish zloty. And finally, point five: like Tom mentioned and illustrated in the extra materials we shared, the forward thrust in our operations is being obscured by the difficulties confronting our biggest client along with significant currency fluctuations. If we were to strip away the impact of that client and those FX movements, our top-line expansion rate for 2024 would have accelerated compared with 2023. With that, let's turn to our outlook.
Edward McGowan
executiveLooking at the first quarter of 2025, our projection for revenue falls between $1 billion and $1.02 billion. That represents a rise of 1% to 3% as reported, or 3% to 5% in constant currency, compared with Q1 2024. Versus Q4, we anticipate Q1 revenue will come in somewhat lower, and here's why: less revenue from our biggest customer, which we already talked about; foreign exchange working against us; onetime license revenue declining in Q1 relative to Q4; Q1 having 2 fewer calendar days than Q4, meaning 2 fewer days of usage revenue; and lastly, quite a few delivery customers hold contracts with calendar year-end dates. So when you add up all those renewals, the combined effect typically drags Q1 revenue below Q4 levels. Keep in mind we believe a full quarter's contribution from Edgio will partly counterbalance these items. Assuming spot rates stay where they are, currency movements should hurt Q1 by $7 million versus Q4, and by $15 million on a year-over-year basis. Given those revenue levels, cash gross margin should land at roughly 72%. For Q1, non-GAAP operating expenses are forecast at $310 million to $316 million; EBITDA margin for Q1 should be around 41%. Non-GAAP depreciation expense should come in at $132 million to $134 million. Non-GAAP operating margin should be roughly 28%. Taking the revenue and spending picture I just described together, Q1 non-GAAP EPS should fall between $1.54 and $1.59, which is down 6% to 3% as reported and down 2% to up 1% in constant currency; this EPS guidance bakes in taxes of $57 million to $59 million, using an estimated quarterly non-GAAP tax rate of about 19.5%; it also assumes a fully diluted share count of roughly 152 million shares. Turning to CapEx. We plan to invest somewhere around $237 million to $245 million in the first quarter, not counting equity compensation and capitalized interest. That works out to roughly 24% of total revenue. As I mentioned before, we expect CapEx spending to be heavily weighted toward the front of the year, with Q1 carrying significantly higher expenditures than the remaining quarters. Now for the full year. We're expecting revenue between $4 billion and $4.2 billion, which is flat to up 5% as reported and up 1% to 6% in constant currency. The range we're giving is wider than what we've done historically, reflecting the bigger scale of our business, uncertainty surrounding our largest customer, and ongoing macroeconomic and geopolitical conditions. If the U.S. dollar weakens significantly, traffic growth comes in well above where we are now, and there's no ban in the U.S. affecting our largest customer, we'd expect to land at the upper end of that range; if instead the U.S. dollar strengthens considerably, traffic slows meaningfully from current levels, and our largest customer does get banned in the U.S., we'd expect to land at the lower end. Now turning to security. For 2025, we're looking at security revenue growth of about 10% in constant currency; within Q4, the combined ARR for our Zero Trust enterprise and API Security solutions came in at roughly $247 million. For 2025, we anticipate that combined ARR from these 2 products will rise 30% to 35% year-over-year in constant currency. On the compute side, we're expecting revenue growth of around 15% in constant currency. Within compute, our cloud infrastructure services finished 2024 with $259 million in ARR. For 2025, we expect cloud infrastructure services ARR to grow year-over-year in the 40% to 45% range in constant currency. Based on current spot rates, our guidance bakes in a negative $38 million year-over-year revenue impact from foreign exchange in 2025. Shifting to operating margins. For 2025, at today's FX rates we're estimating a non-GAAP operating margin of roughly 28%. The operating margin decline across full year '25 stems primarily from depreciation expense running about 90 basis points above last year, along with the year-over-year negative foreign exchange effect. We expect full year capital expenditures to come in at roughly 19% of total revenue, which is up about 2 points from 2024 because of the items I mentioned earlier. As a share of total revenue, our 2025 CapEx should break down roughly like this: about 4% for delivery and security, about 6% for compute, about 8% for capitalized software, with the remainder going to IT and facilities related spending. Turning to EPS. For full year 2025, non-GAAP earnings per diluted share should land in the range of $6 to $6.40. This non-GAAP earnings guidance assumes a non-GAAP effective tax rate of approximately 19.5% and a fully diluted share count of roughly 152 million shares. Before I wrap up, I wanted to touch on our revenue and profitability goals for the next 3 to 5 years. Cloud infrastructure services is seeing significant growth, and we believe it will be the primary driver of our total compute revenue in the near to medium term. Our goal is to grow total compute revenue at a compounded annual growth rate of approximately 20% over the next 3 to 5 years. Now to security. Given the size and scale of our security business, it's only natural that certain products will reach maturity and see slowing growth rates, while others in emerging areas will grow more rapidly. Consequently, we expect security growth over the next 3 to 5 years to be led mainly by our newer Guardicore platform and API Security products; however, some of our largest security products, such as Web Application Firewall or WAF and Prolexic, are examples of our more mature offerings. We've been offering these products for over a decade and have reached a very high penetration rate within our customer base. So because of that, we see the expansion pace for those offerings starting to ease in 2025 and then continuing to taper for the rest of the decade. With those forces at work in our security portfolio, what we're aiming for is to bring total security revenue up at a compounded annual growth rate of roughly 10% across the coming 3 to 5 years, and that figure includes the M&A activity we normally do. On the delivery side, we expect the declines to soften in 2025, with more stabilization anticipated as the next few years play out; to sum it up, we believe that combining double-digit growth in our security products, fast expansion in our compute services, a delivery business that's steadying, and a product mix that keeps shifting away from delivery will get us to double-digit top line revenue growth by the close of the decade, or possibly earlier. Turning to profitability, we're looking for operating margins to hit 30% plus by the end of the decade; and as the top line speeds up and operating margins widen, our objective would be to turn in strong growth in non-GAAP EPS over the next few years. Last, though we aren't giving specific guidance for 2026, we do expect revenue acceleration and margin expansion relative to 2025 levels. That's all from me, and Tom and I are glad to take your questions. Operator?
Roger Boyd
analystNice work on the quarter. Also great news about that big cloud agreement. Tom, I'd appreciate it if you might give us a bit more color there — perhaps share further specifics regarding what kinds of workloads these customers are migrating. It sounds like this is likely shifting away from a hyperscaler, yet anything else you can tell us about how the competition looks when it comes to pricing, performance, and edge coverage would help. I believe Ed mentioned that this customer wants certain conditions tied to specific geographies. That's quite a few questions, I realize.
F. Leighton
executiveRight. For multiple media applications, they've already been consuming a significant amount of cloud infrastructure services from Akamai, and that consumption is going to grow. Beyond that, they have particular requirements in Scandinavia, so for them we're constructing a data center in that region to support major applications across Europe. Thus, in the ordinary sense they're utilizing our services, and on top of that there's a unique case where we're handling a build-out and development effort on their behalf that we expect will eventually be leveraged by a large number of our other customers.
Madeline Brooks
analystTo start, just a quick housekeeping item. I'd like to confirm that the $100 million compute agreement isn't the same thing as that 5-year contract you closed with your biggest delivery client? Then I have one short follow-up on that.
F. Leighton
executiveIt's the same customer.
Madeline Brooks
analystUnderstood. All right, and then the second thing would be this: could you provide additional specifics beyond merely the geographic area and the reasoning behind their selection of Akamai? Might it relate to, say, more recent offerings you've developed, efforts involving AI or intelligence? Anything along those lines would be useful.
F. Leighton
executiveSure — superior performance along with more attractive pricing is why they picked Akamai. Our connectivity across Europe is extremely strong. No other company possesses the presence footprint that's ours. As a result, we're able to serve this client's requests across the whole continent and route them into the data center site they want located in Scandanavia, accomplishing that quite efficiently at a good price thanks to the platform and capabilities we already have.
Amit Daryanani
analystI think I've got two quick questions, hopefully. First, regarding your biggest customer and that $60 million headwind you mentioned, I'm hoping you could maybe elaborate on it a bit. Will that headwind occur no matter what happens with the recent executive order concerning their operations? I'm just trying to figure out whether this is more about them planning to bring some of the CDN in-house, or whether it's more tied to an executive order — does that $6 million headwind happen either way? And it sounds like there's another $60 million in '26. Maybe just give us a baseline on what's going on there and how to ring-fence it — that would be helpful. Then, on operating margins: what do you think it takes for Akamai to return to a 30% operating margin number over time? Any help on revenue or cost control would be helpful.
F. Leighton
executiveYes; so the headwind is mostly due to their DIY build-out, as we've discussed; they'll be taking on more of the services themselves. And there are political challenges, obviously, in the U.S. as a result of the 5-year agreement. Our exposure there is mitigated quite a bit. As we discussed in the disclosure last year, we had about $50 million, have about $50 million of U.S. business with this customer that obviously would be taken away if they are banned in the U.S; however, as a result of the minimum commitments made, our exposure there is less. As Ed pointed out, if they are banned in the U.S; then we would have less revenue from them than if they continue to operate; and if they continue to operate, there would be upside to the upper end of the range; and Ed, do you want to talk about the margins?
Edward McGowan
executiveAbsolutely. So the growth we're looking at will mainly be driven by a couple of factors. First, our mix will shift increasingly toward security and compute. And second, simply in terms of operating leverage, as revenue rises, we're confident we can grow the business in line with that. A significant portion of that extra growth should flow down to the bottom line. Regarding the go-to-market changes Tom discussed or brought up during the call, part of those has already been covered by steps we took earlier, and we'll be very disciplined about any investments going forward; so we anticipate realizing some scale as our top line expands.
Fatima Boolani
analystTom or Ed, jump in whenever you like on this one. About TikTok, you brought up the delivery side of things and those minimal commitments — the minimums they've locked in to shield you. But what I really wanted to know is whether you could share a view or weigh in on whether that customer has picked up other offerings in your portfolio, meaning, say, anything on the security side that we should watch for getting chipped away. And I've got a quick follow-up about the good market too.
F. Leighton
executiveYeah. The way any big customer behaves, they essentially take advantage of every service we offer. So they're a big delivery customer. They lean on plenty of our security offerings and, in fact, an expanding share of our compute — our cloud infrastructure services. And that's already baked into the guidance we provided, covering both this year's guidance and the 3- to 5-year guidance.
Fatima Boolani
analystThanks. From the standpoint of how you go to market, Tom, you made it quite clear that quite a bit of deliberate change has occurred, and perhaps something close to a rebuild of the go-to-market function through things like account segmentation and pricing approach, among others. How much further do you still have to go in solidifying or locking down many of those changes, since from an overall go-to-market strategy angle, a few items appear to still be shifting? So I'd really like to hear where you stand on the effort to rebuild that organization, and what we ought to expect once you reach the other side.
F. Leighton
executiveYes. Excellent question. My view is that this is approximately a 2-year effort. We're at the early part of it; our hunter headcount is already up, and as I mentioned, so is the hunter-to-farmer ratio. We have already grown the specialized sales team and will grow it more. Right now we're at work thinking about changes to pricing structure, contracts, possible changes to contract structure, and making headway on segmentation for the field; so plenty of effort remains in progress. And I think you should frame this against the fact that the company is being transformed; consider our product lineup, which is shifting dramatically from a delivery business, and afterward a delivery plus stadium web app firewall business, where the bulk of our revenue has come from. Then consider the products that are expanding strongly and carrying us toward the future. You're looking at some intake are core segmentation and enterprise Zero Trust security, you're looking at API Security, an enormous greenfield opportunity for us; and finest of all, you're looking at our cloud infrastructure services, which have simply soared. It's striking, I think, once you see what has taken place. Linode was acquired roughly 3 years back, at approximately $127 million ARR, expanding in the mid-teens at best. Now it's more than double, having grown 35% last year, and our forecast is 40% to 45% for this year; and what I find truly thrilling is that our addressable market is literally beyond $100 billion. And that's likely only within our present customer base. So enormous potential, and it's genuinely soaring at this point. So that reshapes the company. It also reshapes what we require from the sales force and what they ought to be doing. Naturally, we intend to safeguard and expand revenue within our current base, but we also intend to be out selling those newer products, both into that base and to plenty of new customers. And so, in response, we are transforming the sales force to build the capabilities required to sustain the very strong growth we're witnessing in these newer areas.
Frank Louthan
analystFrank, can you provide some insight into how logo growth looked during the quarter? That would be helpful. Also, was there a specific area — security, compute, or something else — that experienced stronger or weaker growth from new logos?
Edward McGowan
executiveFrank, it's Ed here. The gross logo figure, frankly, was likely helped the most by those Edgio contracts we brought on board. A few hundred logos came with that, and more than 100 of them were brand new to us. That made up the largest piece of the increase. Regarding the source of these new logos, they're mainly from security, specifically within the 2 fresh areas Tom mentioned — Guardicore and API Security — plus the compute segment too. So compared to a typical acquisition quarter, Q4 was probably above average for new additions, just through regular business, and clearly Edgio contributed significantly on top of that.
Frank Louthan
analystSo how do you see the situation with keeping those Edgio contracts? Are some likely to drop away? Or how probable is it that they turn into longer-term accounts?
Edward McGowan
executiveRight. At this stage, the migration is behind us. Everybody has been moved off. The Edgio network is now closed — or shut down now. So we have a solid sense of who is transitioning over now; from the base we have today, we don't expect meaningful churn. We were extremely picky about what we accepted. Not every customer was taken on. Some things, some smaller customers, simply didn't make sense for us to bring aboard. There are certain acceptable use policy violations we would not accept, and with some customers, the economics weren't something we wanted. So built into our guidance is what we believe we'll achieve with that customer base. My sense is there's likely some upside there through upsell. Broadly, the approach is to land the customer, ensure they're satisfied, and then begin building a pipeline with them — you don't want to immediately hit them with a sales pitch. So we're quite pleased with what we managed to retain; it came in basically right on target, perhaps slightly better than we had anticipated.
Rishi Jaluria
analystI've got 2 fast questions. To start, returning to a few of the shifts in go-to-market and the adjustments on the sales side: what are you doing internally to keep the disruption as small as possible? And what assumptions sit behind your 2025 guidance? The reason I'm asking is that throughout software's history, whenever salespeople depart or go-to-market changes happen, the fallout tends to be far bigger than anyone expected. So if you could take us through what you're assuming and the steps you can take to hold that disruption down, that would be useful. Then I have one quick follow-up.
F. Leighton
executiveWell, the number one priority is preventing an account breakage. That's why we aren't doing everything simultaneously. Like I said, this is a 2-year journey. And naturally, we're bringing in people with expertise around the new products, so those individuals are assisting with cross-selling and get an opportunity to pursue new customers. So I believe the main thing is that we'll keep breakage of rep relationships with customers in the accounts to a minimum. And Ed, would you like to add anything there?
Edward McGowan
executiveAbsolutely. Like Tom said, we've engaged one of the big consulting outfits, bringing in some know-how to assist us in thinking that through. That said, we definitely intend to proceed with a lot of care regarding the shift from hunting to farming and the way we back customers going forward. The lower the breakage, the better.
Rishi Jaluria
analystUnderstood. No, that helps. Then, very briefly regarding Edgio: I value the context and the thorough guidance. When we consider things over a somewhat longer horizon, now that Edgio has been removed from consideration and those agreements and resources are in your hands, what should our view be concerning the pricing landscape for delivery? Also, could the simple fact of there being a smaller number of competitors in this area contribute to a pricing landscape that is less aggressive?
Edward McGowan
executiveYeah. The answer depends greatly on which customer segments you're looking at. For certain of the bigger, high-volume clients, we're observing a degree of rationalization, with extended contractual durations, which is a positive. Clearly, having fewer competitors can mean you potentially get improved pricing, yet that's something we're already beginning to observe in the marketplace. So I would characterize it as improving somewhat, though this has always been an industry propelled by volume. Consequently, if volumes keep pushing upward, we will still witness the overall downward direction in pricing, but with luck, we'll observe it becoming somewhat more tempered going forward.
John DiFucci
analystSo Tom and Ed, I believe this is relevant for both of you. Things seem to be progressing as expected, and we're grateful for those 3- to 5-year targets. Honestly, everything you provided us, that's excellent. However, the outlook for the coming year suggests 3% growth. And I believe there's a 1% currency headwind, which brings it to roughly 4%. I understand the challenges with the large customer, but challenges are always present, correct? Setting that aside for the moment, how can we feel confident that growth will speed up to 10%? Is it purely arithmetic? Because I haven't worked through it yet at all, that the faster-expanding segments of the business will more than compensate for the delivery business? And are you — or perhaps, I suppose, are you anticipating a better macroeconomic environment in the projection? Or does it — I guess, tied to this as well, does that 10% — the eventual 10% leave out any effects from that large customer?
Edward McGowan
executiveAlright. I'll handle this one, Tom. Let me begin with the latter portion. Regarding the buyer's cost, what we've done is we've built a model for what their minimum spend would be under that 5-year contract; so consider that as the floor of what's possible for that customer. So potentially there's additional upside there. But when it comes to what we're observing for growth, I think you need to sort of divide the business into the categories and why we presented all that information to you. Let's begin with delivery. So delivery, clearly, has been a fairly significant drag. It's beginning to get better; we anticipate it will get better this year, and keep getting better. So you ought to begin seeing some stabilization in delivery during that timeframe. So then you consider that as a major drag kind of easing considerably, so you don't face as much of a drag. So then when you look at the businesses we have, consider security, the larger franchises are decelerating somewhat, but they're still expanding. So we're still achieving some growth from that. And below that, say, for instance, in our security products, we have new fraud products and will be launching new products; so that franchise will continue to grow, it just won't grow as quickly as before; but the — between Guardicore and API, those are very sizable markets; we have market-leading products. We believe we'll be able to — that combined with the — some of the adjustments we're making in go-to-market and the investments we're making in hunting, we'll achieve more productivity from the sales force. And there's still a great deal of distance to cover there. So we believe that, that growth as it expands, you get 10% growth from security, which we think is quite sustainable given those factors. And then there's compute. When it comes to compute, we're undergoing something of a shift—Tom mentioned this—involving what I'd call, well, not exactly legacy, but a number of the more mature offerings sitting in that other applications bucket. We had been working alongside certain ISV partners, and that has caused a slight near-term revenue flattening within that bucket. Even so, it opens up a considerably larger chance for us to expand the market through those partnerships, capturing a portion of their own activity while also securing a product that receives ongoing investment, and so forth. As a result, our offering is far stronger; and beyond that, thanks to the go-to-market spending we're directing toward compute—not merely through the hunters, but through the overlay specialists as well—we believe that will further boost productivity out in the field. So I hope that addresses your question.
John DiFucci
analystThat was genuinely impressive, and quite useful. It leaves me feeling there's a great deal more to explore, particularly on the compute side, though thinking about Guardicore and API is thrilling. I suppose there's just one item left — a final query regarding the compute segment. Something I keep coming back to is profitability, and that's a concern raised about anyone entering this space. You mentioned possessing certain benefits when moving into this area because of the delivery operation. Could you flesh that out a bit further? I grasp the broad picture, I think. Does it boil down to leveraging underused resources throughout your network, or is there more going on?
F. Leighton
executiveMultiple factors are at play. To begin with, it's likely we handle more data movement than virtually anyone else, and we've mastered doing so with great efficiency. Consequently, for clients using our cloud infrastructure platform, comparing our rates against the hyperscalers shows ours are considerably lower. At minimum, for the workloads involving data movement, they encounter plenty of hits, and such workloads represent a large share of what exists currently. Thus, in direct competition today with the hyperscalers, we're able to present a far superior price. The second factor is performance. As we've discussed before, ours is the most distributed platform globally, which lets us position our customers' compute instances much nearer to their end users. Beyond aiding scalability, this significantly boosts performance, since business logic executes in proximity to users. Hence, when we run direct trials today versus the hyperscalers, across many applications, our results are stronger. That means we can deliver enhanced performance at a cheaper price. Plus, with the managed container service now available, it's incredibly thrilling. We're capable of hosting our customers' containers across hundreds of cities. No one else offers that currently. So once more — and those cities are places we already serve through the delivery and security operations. We're established there already, so you receive improved performance and scalability at a reduced price. Then you factor in the massive market we can potentially reach. I believe it's truly crucial, as you build your models, to examine our cloud infrastructure services revenue and its trajectory. We only began offering that to enterprises in the prior year and went from minimal revenue to $115 million ARR, with the overall figure for our cloud infrastructure services hitting $259 million. We're forecasting that overall figure to increase by 40%. And at minimum 40% to 45% during this year, so I'd urge you to monitor that closely, since it fuels substantial growth and profitability for us ahead.
Edward McGowan
executiveJohn, to build on what was said about profitability, four key synergies exist. The first one is the backbone. Naturally, we're able to take advantage of it. Clearly, we have agreements we fulfill in that area. Second comes operations. Many of the same people who construct the CDN likewise construct the data centers for our compute business, so bringing on a large team wasn't necessary. Third is engineering. Earlier this year we mentioned that roughly 1,000 individuals were shifted from the engineering organization within our delivery business over to the compute business, across both operations and engineering. The fourth item concerns go to market. Thus, we're able to take advantage of our current customer connections and our reps. We're supplementing that somewhat with certain specialists of ours, yet going out and bringing on an entirely new team isn't required; therefore, a great deal of operating expenses can be utilized by us when it comes to driving profitability.
John DiFucci
analystAll of that tracks. The simple portion comes next. All you need to do is deliver. That was meant as humor. It isn't simple.
Patrick Colville
analystI think this next one is likely aimed at Dr. Tom and Ed. I mean, during the prepared remarks, you addressed what caused compute to be sluggish. But I'm just — right now I'm working through my model, and the year-on-year difference is quite dramatic. So I simply want to confirm I have those reasons fully pinned down; so to begin, I suppose, this year, compute grew 25%, and for next year your guidance is 15%. Could you elaborate on what's behind that?
Edward McGowan
executiveYes, of course. There are 2 components here. One is cloud infrastructure security, which is going to keep expanding and in fact speed up. That component is the smaller of the 2 businesses, yet it's going to accelerate quite substantially. If you dig into the details of the other one, application services, there are several elements involved. For instance, we have a portion of our legacy net storage business, and we'll be end of life-ing that shortly within the next year or 2; and certain customers will move to our cloud offering, others might choose to do something different, but that's going to decline somewhat; and then a significant amount of revenue comes from some of these potential ISV partners. For instance, we have some of our transcoding. We put out an announcement with a partner named Harmonic, and we're teaming up with them while shifting some of that business in their direction. They now rely on us for their compute. So that's going to change gradually. And there are roughly 4 or 5 separate buckets, whether that's image manager, video manager, our waiting room application or visitor prioritization. We're going to move those toward ISV partners while concentrating more of our efforts on the infrastructure services. So you'll notice flattening, perhaps a slight decrease, in that bucket, and the bulk of growth will come from cloud and infrastructure services.
Patrick Colville
analystAlright. Alright. That's useful. Then, just a fast one, I suppose, since this question keeps coming up from investors; regarding Akamaized's exposure to the U.S. Fed. What's the right way to think about that? And for 2024, does guidance for 2025 include any assumptions for that business?
Edward McGowan
executiveSo, when you mention the U.S. study, did you mean the Federal Reserve's interest rate policy? Or are you referring to the federal government? < >.
Patrick Colville
analystThe second one.
Edward McGowan
executiveAll right. Yeah. Clearly, we do make sales to the government; that isn't a massive chunk of what we do. It's far too soon to say, given the cost-cutting moves currently underway, what kind of effect that might have on us. Still, it isn't a particularly significant piece of the business. I wouldn't anticipate anything significant coming from that. Nothing significant has been built into our models — just business as usual at this point. So I'm not expecting anything.
James Fish
analystI appreciate all the information provided, and the expanded disclosures, I believe investors will too. Perhaps just to piggyback on a point you made earlier, Ed. Could you take us through a bit of that transition effect on the compute side and which products you intend to be moving away from, and how should we consider the utilization of the compute network currently that we must invest this heavily after two years of investing so much into the compute business.
Edward McGowan
executiveYes. I mean I provided you regarding the investment percentage of revenue that we're allocating for that. 1% of that out of the 6 that we're utilizing compute will be earmarked specifically for that large customer. And I've consistently mentioned how occasionally, if we do secure big customers, you do tend to sometimes see outsized demand in a particular geography. We put in a tremendous amount of investment this year regarding expanding a number of locations, and so on; we're going to keep adding to those. We're also witnessing very robust demand. So we're continuing to put money in with that. But also remember we did migrate more than $100 million worth of our own applications. It's currently running above $0.5 billion of our revenue, and that business keeps expanding; so we're still having to make some investment to support that, which is much lower than what we would be paying if we were to keep using third-party cloud providers; now regarding the 2 different components of revenue, I think I covered the cloud infrastructure service. So I think you all understand that. That's the genuinely fast-growing part. And within that other bucket is where you'll observe that flattening out to declining. And it's really just the items I referenced. We've got video management, image management, visitor prioritization and some various workflow components such as transcoding that we have built occasionally for customers. There isn't a great deal of demand for it. It's grown reasonably, but it's not a specialty of ours. So we want to ensure that we're investing in the correct areas. So as we're migrating that to partners and we shift our investment into the faster-growing infrastructure cloud services, you'll see a bottoming out here this year and then the growth will reaccelerate to 20%, led by cloud infrastructure services. And also keep in mind, I've got roughly — last time we broke it out, it was around $50 million of legacy storage; so I think of that as similar to the media companies storing like a backup origin for their video files and music file and that kind of stuff, images, and so on. That's — that business will be winding down over the next year or so. So that's going to have a bit of a drag on revenue as well.
Patrick Colville
analystGot it. And I believe plenty of investors on this call — you're here presenting a 3- to 5-year framework. The previous occasion you provided us with a framework like this was quite a few years back; and my recollection is that within it, we discussed something like 20% constant currency security growth, and we fell short of that. So now you're referring to a 10% CAGR that incorporates M&As from this point within security. What exactly is going to be the driver behind us truly delivering these results, and what makes this a sensible expectation — one that could even prove cautious rather than us simply hoping we reach that figure?
Edward McGowan
executiveWell, to be honest, across that stretch there were multiple years where growth of 20% did happen. We consistently noted that 20% would involve acquisitions. So along the way we made some acquisitions. We've — look, we just recorded 16%. So that business has been quite robust. For years now we've been bringing in a couple of hundred million of high-margin security revenue. And if you run the numbers, that's going to keep happening. Regarding how acquisitions will factor moving forward, we don't expect to make significantly large ones, more along the lines of what we've historically done, be it tuck-ins or smaller businesses that have — possess a product already in market that's nearing scale, such as API Security, where the company is contemplating a major go-to-market investment or a raise or an exit. We acquire something of that nature, or like Guardicore, and we scale it up, and we've demonstrated success in that regard; so I'd say our confidence in the M&A strategy is solid. It won't represent a significant portion, but perhaps as time goes on, it could be among the faster-growing products by the time we exit the decade. And the reason we've — why we broke these businesses out is so you can observe how effectively we're now scaling a business that's a couple of hundred million and growing extremely fast. So I'd say our confidence is pretty good. And once more, yes, you can — we can argue about the 20%, and perhaps for a couple of years we're in high teens. But as we stated, acquisitions would have been part of that and we always do them, yet I believe we performed pretty well across that stretch. And we simply felt that now, with us reaching a scale of a couple of billion, updating expectations made sense as some of our products that have been on the market for over a decade are slowing down.
Will Power
analystTom, during your prepared remarks you gave several solid illustrations, I believe, of a few of the applications running across your network. For a while now it has appeared that you ought to be nicely situated when it comes to inference at the edge. And I suppose — I'm curious whether you might give an update on that, simply regarding the character of the discussions you're having, how the pipeline is shaping up, and the level of activity you're observing when it comes to capitalizing on especially a few of these generative AI applications and inference that — perhaps it's beginning to emerge, or at minimum you believe it will begin to emerge?
F. Leighton
executiveYes, quite a bit, in my view. AI-driven image normalization, so that the different images share one uniform size and quality, image classification, even for tasks such as spotting disease in crops. AI-driven speech recognition inside cars, AI-driven chatbots, sentiment analysis, text-to-image, inferencing across a range of applications. Every kind of application already putting Gen AI to use on the platform. And we have partners, our ISV partners. A number of them are doing AI on the platform, web assembly, AI inferencing. So I'd describe it as early days, yet every kind of thing is being done today on our cloud infrastructure services.
Will Power
analystOkay. Turning to the acceleration we're assuming in cloud infrastructure — I believe you're targeting 40% to 45% ARR growth. Can you break that down further? How much of it is inferencing driving? Or is that still early? And if not, what are the main drivers behind that acceleration, including the effect of the $100 million deal — how does that fit in?
F. Leighton
executiveYes. That $100 million deal contributes little in the current year; it really only ramps toward the end of the year. So my view is it won't materially affect our 40% to 45% projection. It will definitely be a help in '26 and beyond. And in my view AI inferencing isn't a major contributor to this year's growth either. There may be a bit, but I'd put that more in the upside bucket. I think it's mostly the traditional uses of compute on the platform; and remember, that's a $100 billion market today, plus — without factoring in AI or the new large deal we've done. Just to give examples, with our ISV partners: database at the edge, observability for all sorts of applications, media workflow, things such as file transcoding and live encoding, video packaging, interactive WebRTC; those are what our ISV partners have their applications doing on our cloud infrastructure platform — digital asset management, video optimization, game orchestration, fleet management, DRM, Kubernetes activity and auto scaling, ad insertion, service side ad insertion. So there's simply a bunch of stuff that is just good old regular compute running on the platform. I'd say we're — as the examples we've given show, we're selling it into pretty much every vertical. A lot of financial companies are beginning to use it now. Media, broadly construed, is still, in my view, the sweet spot, and partly that's intentional, because those customers need to get great performance. They need to cut cost. They don't like writing a huge check to their biggest competitor. And they're big Akamai customers, and they like Akamai a lot; so that's sort of the center of mass, but it goes across verticals and across a lot of use cases. And I'd say we're excited about AI. That's upside and probably comes down the road.
Mark Murphy
analystTom, I'd love to hear how you think Akamai gets affected by everything that's been happening with the DeepSeek model in terms of tech progress and efficiency gains. Also, are you noticing anything that suggests this is kicking off a wider period of people trying out edge inferencing on Akamai's compute? Perhaps a number of these AI use cases will start to make more financial sense; and then I've got a quick follow-up for Ed.
F. Leighton
executiveSure. My take is that this DeepSeek phenomenon, announcement, whatever label you want to put on it, lines up really well with the predictions we've been making. I expect you'll keep seeing more gains. For us it's a positive, since it confirms both our prior statements and the way we built our cloud infrastructure platform; as a matter of fact, some organizations are already using DeepSeek on that platform, alongside other models. And what it does mean is that the cost drops considerably; you don't require those massive core GPUs. Lighter weight GPUs are enough to run it. Plenty of scenarios exist where having that at the edge is what you'd want. To me it's excellent and basically aligns with what we anticipated; and I believe you'll witness additional progress in that same direction.
Mark Murphy
analystAll right, that's wonderful to learn. Then, Ed, following up regarding the security CAGR, for which you anticipate roughly 10% over the coming 3 to 5 years — that seems pretty near, I think, to the broader market expanding around 12% to 14%; plus you mentioned it factors in a normal amount of M&A. If we look at it organically, ought we to model something akin to perhaps mid- to high single digits on an organic basis? And afterward, if you can locate a few of those tuck-ins — you've performed amazingly at selecting them and carrying them out lately, and they're expanding like crazy — then perhaps the inorganic portion hands you a few points there across the next 3 to 5 years. Does that work as a reasonable framework?
Edward McGowan
executiveYeah, I'd say that's a solid approach to consider it, Mark. Clearly, given the scale we now have, above $2 billion, and we discuss how even the larger revenue companies we bought were $20 million to $30 million — that's negligible, roughly a point overall from inorganic. Regarding the near term, it will come mostly from organic; yet yes, likely around 1 point, something along those lines once you get a couple of years out. Then clearly, if we selected a strong company as we did with Noname or as we did with Guardicore, as you move further out, we treat that as organic growth by then. That could be adding since it's a faster-growing area if we enter a hot space expanding rapidly. But regarding the — to hit a figure in any particular year, we don't expect to get much beyond roughly 1% from an acquisition inorganically.
Jonathan Ho
analystGreat, and I'll add my thanks for the extra details. I've just got one item. How worried ought we to be about possible tariff effects? And is it feasible for you to hand along the increased expenses tied to that? I realize you're perhaps looking to speed up certain investments early, especially since you need to put money into the compute area over the extended term; it appears there could be a bit of risk here, so I merely wanted to grasp what that means.
Edward McGowan
executiveYeah. Clearly it's hard to predict, simply because we don't know what the final outcome will be regarding towers. Yet if we can relocate supply chains, and we're clearly considering that. There was a bit of discussion regarding Canada and Mexico. We do obtain a few server builds from that region. But — so let's jump ahead a bit — I provided the figure, $10 million to $15 million; thus it isn't major. As for whether we can shift some of those price increases onward, we're definitely looking into that as part of the effort we're doing with the consulting firm we brought on. The whole pricing approach is included. And if it turns out to be rather inconsequential in some manner, we'd definitely need to incorporate that into our pricing. Still, it's hard to determine right now since we simply don't know what the ultimate end state will be.
Mark Stoutenberg
executiveAppreciate it, everybody. Just like always, we'll be showing up at investor conferences for the remainder of this quarter. We're eager to meet you at those events and go over all the topics covered during today's session; so once more, thanks for being with us this evening. I realize this has been a lengthy call; we hope the rest of your eating — your evening is a pleasant one. Operator, please go ahead and close out the call now.
Operator
operatorThat wraps up this conference. We appreciate your participation in today's presentation. You're welcome to disconnect at this time.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Akamai Technologies, Inc. transcript — plus 133,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.
Get the API View API docs →This call discussed
For developers and AI pipelines
Programmatic access to Akamai Technologies, Inc. earnings transcripts and 133,000+ others is available through the
EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments,
full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.
More tools for stock research
OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Akamai Technologies, Inc. or any ticker on your watchlist.
StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.