Akamai Technologies, Inc. (AKAM) Earnings Call Transcript & Summary

August 6, 2026

NASDAQ US Information Technology IT Services earnings 59 min

What were the key takeaways from Akamai Technologies, Inc.'s August 6, 2026 earnings call?

In Q2 2026, Akamai Technologies reported revenue of $1.1 billion, reflecting a 5% year-over-year increase, while non-GAAP earnings per share (EPS) came in at $1.59, down 8% year-over-year. The company announced a significant $600 million, four-year commitment from a U.S.-based technology company for cloud infrastructure services, contributing to a total of over $2.8 billion in multiyear commitments for the year. Management signaled an acceleration in overall revenue growth into the low teens for 2027, supported by a strong pipeline and strategic investments in cloud infrastructure and security solutions.

What topics did Akamai Technologies, Inc. cover?

  • Revenue Growth Acceleration: Akamai anticipates overall revenue growth to accelerate into the low teens in 2027, driven by strong multiyear commitments and a robust pipeline. CEO Tom Leighton stated, "we now anticipate that Akamai's overall revenue growth will accelerate into the low teens in 2027."
  • Cloud Infrastructure Services (CIS) Performance: CIS revenue reached $99 million, up 39% year-over-year, indicating strong demand and growth potential. CFO Ed McGowan noted, "we continue to expect CIS revenue to meaningfully accelerate in Q4 and then further accelerate in 2027."
  • Security Revenue Growth: Security revenue grew 10% year-over-year, driven by demand for web application firewall and API security solutions. McGowan highlighted that "Security revenue maintained strong momentum with revenue of $604 million, up 10% year-over-year as reported and 9% in constant currency."
  • Acquisition of LayerX: Akamai completed the acquisition of LayerX for approximately $205 million, enhancing its security portfolio. McGowan mentioned, "we expect LayerX will be dilutive to our non-GAAP EPS by approximately $0.12 for 2026, split evenly across Q3 and Q4."
  • CapEx and Growth Investments: Akamai plans to invest significantly in capital expenditures, projecting $475 million to $525 million in Q3 to support its growing CIS pipeline. McGowan stated, "we expect to invest up to $500 million in CapEx to replenish and expand our GPU capacity."

What were Akamai Technologies, Inc.'s August 6, 2026 results?

  • Revenue: $1.1 billion (+5% YoY)
  • CIS Revenue: $99 million (+39% YoY)
  • Security Revenue: $604 million (+10% YoY)
  • Non-GAAP EPS: $1.59 (-8% YoY)
  • Operating Margin: 25% (inline)
  • CapEx: $347 million (32% of revenue)

Akamai's strong performance in Q2 2026, highlighted by significant customer wins and robust revenue growth in cloud infrastructure and security, positions the company favorably for future growth. However, analysts are cautious about potential margin pressures and the timing of revenue recognition for large contracts. Investors should monitor the execution of growth strategies and the impact of macroeconomic factors on performance.

Earnings Call Speaker Segments

Mark Stoutenberg

executive
#1

Thanks for being with us this afternoon, everybody, and welcome to the earnings call for Akamai's second quarter of 2026. Tom Leighton, who serves as our Chief Executive Officer, and Ed McGowan, our Chief Financial Officer, are the speakers you'll hear from today. I'd like to point out that revenue and earnings guidance are among the forward-looking statements made in today's remarks. Current expectations and assumptions form the basis for these forward-looking statements, which are subject to certain risks and uncertainties and involve a number of factors that could cause actual results to differ materially from those expressed or implied. Macroeconomic trends, the integration of any acquisition, geopolitical developments, and other risk factors identified with our filings with the SEC are among those factors, though the list is not limited to them. The statements made on today's call reflect the company's views as of August 6, 2026, and no obligation is assumed by us to update any forward-looking statements. As a reminder, certain non-GAAP financial metrics will be referred to during today's call. In the Investor Relations section of akamai.com under Financials, you can find a detailed GAAP to non-GAAP reconciliation. With that, I'll now turn things over to our CEO, Dr. Tom Leighton.

F. Leighton

executive
#2

Thanks, Mark. It's a real pleasure to share that Akamai keeps gaining steam and earning broad recognition across the industry as a critical infrastructure partner for the AI-driven economy. For a long time, Akamai has been recognized for running the planet's most distributed platform for content delivery and cybersecurity on a global scale, with a reputation built on reliability, quality and trust; and at this point we're putting our worldwide footprint and many years of experience serving the biggest enterprises on earth to work, doing for the AI-driven economy what we've already done in cybersecurity and content delivery. The feedback from industry leaders at major enterprises has been extremely positive. In fact, we were delighted to reveal today that a U.S.-based technology company has pledged over $600 million across 4 years for our cloud infrastructure services to support their robotics development. With this new customer added, the overall volume of multiyear commitments signed during this year for our cloud infrastructure services now exceeds $2.8 billion. Given these commitments and the exceptionally strong pipeline in front of us, we now expect that Akamai's total revenue growth will speed up into the low teens in 2027. For investors eager to grasp how the AI market is developing and why Akamai holds such a unique and essential position in this new ecosystem, we suggest you look at the July 23 article in Fast Company written by Victor Dey; there, Dey describes the part Akamai's distributed platform has played in reshaping the content delivery and cybersecurity marketplaces, and how a comparable approach could serve the AI ecosystem, with many Agentic workloads handled at the edge, near users rather than inside huge centralized data centers. In a blog tied to that piece, IDC analyst Dave McCarthy examines a world reshaped by AI agents and the real-world difficulties of automatically executing queries against a trained model. While training frontier models demands enormous data centers and heavy energy use, the following hurdle for AI is what's required to operate those models everywhere with low latency and at a cost that's affordable; it's also worth reading a recent blog from Akamai's CTO, Bobby Blumofe, about how AI inference is remaking the cloud. Bobby's post lays out why the infrastructure AI agents need will end up being a flexible continuum, reaching from the core out to the edge. As the AI market advances past centralized AI factories, Akamai's approach is to deliver a unified distributed grid for AI inference. By moving AI inference out to the edge and pairing it with our enormous deployment of CPUs for delivery, security and functions as a service, our platform will let customers run agents and models within milliseconds of their end users, with the responsiveness of local compute and the scale of the global web, optimizing performance while cutting latency and cost. Customers are already using our cloud infrastructure services across a wide range of repeatable use cases. Take, for instance, a company in Singapore offering an AI patent intelligence platform — they moved their inference workloads off a hyperscaler and onto Akamai, which lifted performance by 30% while reducing infrastructure costs by 20%. Then there's an AI-driven decision intelligence platform in Poland that picked us to handle low-latency game theory simulations across the globe, with costs they can predict. A U.S. enterprise data storage provider selected us to run their platform for high-throughput storage observability and analytics. In India, a global AI software company picked us to run their Gen AI image creation — a $12 million win taken from a hyperscaler. Also in India, an AI-powered communications provider selected us to do away with unpredictable hyperscaler billing and speed up their customer engagement. And a SaaS media workflow platform in the U.S. picked us for high-throughput live media and coding. When we step back and look at the whole business, AI isn't just fueling adoption and revenue growth for Akamai Cloud — it has also served as a tailwind for our security solutions. During Q2, Akamai Security revenue rose 10% year-over-year as reported and 9% in constant currency. Once again, security growth was driven by robust demand for our market-leading web app firewall, API security and Guardicore Segmentation solutions. Demand for our WAF stayed strong in Q2, coming from customers wanting protection against vulnerabilities that frontier models such as Mythos or GPT-5.5 could expose. Among those who added or substantially increased their WAF usage in Q2 were one of the world's leading commerce sites, one of the world's leading automakers, and one of the world's largest banks — the latter via a $14 million upgrade to have Akamai secure every one of their applications. We also closed renewal upgrades covering security and CIS products with one of the world's largest telcos, a contract valued at more than $20 million across 2 years; and today we're delighted to share that CrowdStrike, a leading AI-native cybersecurity platform, has moved to Akamai for its web security and content delivery needs. According to CrowdStrike, they were unhappy with the inconsistent service coming from one of our SMB-focused competitors. We view this validation from yet another security leader as a powerful endorsement of Akamai's enterprise security capabilities, along with our hard-won reputation for trust, reliability and dedicated customer support; on the go-to-market side, we announced recently that Akamai has been chosen as a strategic partner for WWT's AI Readiness Model for Operational Resilience, or ARMOR as it's known for short; it's the industry's first holistic vendor-agnostic AI security framework, and having us included establishes Akamai as a foundational security architecture for the AI factories WWT is building and NVIDIA is accelerating. Absent such a collaborative security framework for AI, organizations frequently end up piecing together fragmented security strategies. By matching the Akamai security portfolio to ARMOR, we deliver a methodology for protecting large-scale AI clusters proactively by stopping the lateral movement of threats. In the words of WWT's Chris Konrad, "No single vendor can secure the AI frontier alone. Working hand in hand with Akamai, we are transforming what many treat as mere buzz around secure enterprise AI into something customers can actually deploy at scale. Akamai also takes pride in ranking among the security vendors the industry considers essential, and we partner with Anthropic and OpenAI to support safe, swift rollouts of AI-powered defenses. Thanks to our access to initiatives such as Glasswing and Daybreak, plus our involvement in efforts like the Open Secure AI Alliance, we bring our know-how to bear in keeping large enterprises and critical infrastructure protected. As staff throughout every sector reach for AI tools more often, organizations require greater assistance to safeguard their people and stop confidential information from escaping; that is the reason we bought LayerX, a top supplier of secure enterprise browser and AI usage control; through LayerX, which now carries the name Akamai Workforce Protector, we can hand security teams the deep visibility required to observe precisely how users engage with web content, SaaS applications, file uploads and even raw AI prompts. With this browser security platform folded into our offerings, Akamai stands in a uniquely strong position to shield enterprises from unauthorized AI agents, whether those agents run on someone's laptop or reside within internal applications that communicate with external LLMs. What makes this pairing especially thrilling is the smooth fit between Workforce Protector and the Zero Trust portfolio Akamai already offers, among them our Guardicore microsegmentation, Zero Trust Network Access and DNS security solutions that thousands of enterprises depend upon today. In the end, we think these merged capabilities will produce a potent, unified workforce security offering that tackles head-on one of the sector's most pressing problems: securing and governing the way employees, partners and supply chain ecosystems engage with AI. LayerX represents the most recent of several acquisitions we have completed to expand our security portfolio and to make it simpler for enterprises to secure broader solutions from Akamai as their strategic security partner. All told, we anticipate our security portfolio will bring in over $2.4 billion in revenue during this year, which makes Akamai one of the biggest security providers on the market. To wrap up, we are delighted to watch AI fuel demand for our Cloud and Security portfolios, and we feel thankful for our gifted team here at Akamai, whose extraordinary efforts continue to produce flawless digital experiences for our customers. Our employees' commitment to our customers' success helps explain why so many of the world's top brands and most demanding businesses count on Akamai as their trusted, dependable and reliable partner. I will now hand the call to Ed for more on our results and our outlook for the rest of the year. Ed?

Edward McGowan

executive
#3

Thanks, Tom. Before I walk through the usual quarterly financial review and our updated 2026 outlook, I'd like to add to what Tom said. Strategically and financially, there's a lot for us to be enthusiastic about when it comes to where our business is heading. We locked in a GPU services agreement worth $600 million over 4 years with a top U.S.-based technology company that concentrates on robotics development, and that pushes the major wins we've announced recently past $2.8 billion for this year — a tremendous endorsement of our platform. On top of delivering excellent multiyear revenue visibility, it highlights just how large our cloud infrastructure services business has become. With this new customer, we don't anticipate any material revenue effect during 2026, though we do anticipate revenue ramping completely across 2027. And lastly, as Tom noted before, and supported by these multiyear commitments together with a pipeline that keeps expanding, we can clearly see our way to speeding up top line revenue growth from single digits this year to the low teens in 2027. So with that, let's turn to the Q2 results. Second quarter revenue came in at $1.1 billion, an increase of 5% year-over-year as reported and in constant currency. Cloud Infrastructure Services, or CIS, revenue reached $99 million, rising 39% year-over-year as reported and in constant currency; that matched what we expected, and we still expect CIS revenue to meaningfully accelerate in Q4 and then accelerate further in 2027. Security revenue kept up its strong momentum, posting $604 million, which is up 10% year-over-year as reported and 9% in constant currency. Our fast-growing API security and Guardicore Segmentation solutions remained the drivers of that second quarter strength. Turning to Delivery and other cloud applications. Revenue was $396 million, a decline of 6% year-over-year as reported and a decline of 5% in constant currency. International revenue totaled $549 million, up 6% year-over-year, or up 7% in constant currency, which works out to roughly 50% of total revenue in Q2. U.S. foreign exchange fluctuations hurt revenue by $2 million on a sequential basis and by $1 million on a year-over-year basis. Now on to profitability. In Q2, we produced non-GAAP net income of $236 million, or $1.59 of earnings per diluted share, down 8% year-over-year as reported and down 6% in constant currency. These results reflect our expanded colocation investments, higher depreciation and increased headcount costs, all of which are meant to help fuel faster growth in our cloud infrastructure services. Our non-GAAP operating margin for Q2 was 25%, in line with our expectations. We expect operating margin to stay in the mid-20s for the rest of this year as we keep investing to capture the exciting growth opportunities in CIS. Our Q2 CapEx was $347 million, or 32% of revenue. Second quarter CapEx came in below our guidance, mainly because of the timing of receipt of some GPUs. Those shipments arrived a few weeks later than expected following the quarter end, which pushed the bulk of the planned spend into Q3. Turning to cash and how we're allocating capital: during May, we brought in $3.5 billion through 2 equal tranches of convertible debt carrying a 0 coupon and maturing in 2030 and 2032, with the proceeds meant to fund our expanding CIS pipeline and for general corporate purposes. In Q2, we laid out roughly $410 million repurchasing about 3 million shares. So far this year, we've bought back around 5 million shares, spending a total of about $616 million. We closed the second quarter with about $565 million left under our current repurchase authorization. That said, because market demand for CIS has been so strong, we're hitting pause on buybacks for now so we can redirect capital toward supporting our high-growth CIS pipeline. And lastly, at June 30, our cash, cash equivalents and marketable securities totaled roughly $4.6 billion. Before walking through Q3 and full year 2026 guidance, I want to address some housekeeping items. First, on July 2, we closed our acquisition of a security company called LayerX for about $205 million. Financially speaking, we don't expect this deal to materially affect full year 2026 revenue. On the bottom line, we anticipate LayerX will dilute our non-GAAP EPS by roughly $0.12 for 2026, divided evenly between Q3 and Q4. Second, let me discuss foreign exchange and how it should affect our second half 2026 performance. Given currency movements since our last earnings call, we now anticipate second half revenue headwinds of about $9 million. Finally, I want to give some additional color on capital expenditures for the rest of the year. For Q3, CapEx should step up considerably, partly because of the shipments that slipped into July that I mentioned earlier, and as we ramp up build-outs supporting the major contracts announced earlier this year. As we said on our last call, GPU demand is still exceptionally strong. Consequently, our GPU capacity is entirely sold out. So, driven by the $600 million new customer win we announced today and a very robust pipeline, we plan to invest up to $500 million in CapEx to replenish and expand our GPU capacity. Of that, we expect to spend approximately $60 million this year, with the remainder landing in early 2027. Now on to guidance. For the third quarter, we project revenue between $1.105 billion and $1.13 billion, representing growth of 5% to 7% as reported and 5% to 8% in constant currency versus Q3 2025. At current spot rates, foreign exchange fluctuations should negatively impact Q3 revenue by $2 million compared to Q2 levels and by $8 million year-over-year. At these revenue levels, we expect cash gross margin of approximately 70%. As a reminder, gross margin is affected by the significant increase in colocation as we accelerate growth in our CIS business. Q3 non-GAAP operating expenses are projected at $347 million to $359 million. We anticipate Q3 EBITDA margin of approximately 38% to 40%. Our projection for non-GAAP depreciation expense is $153 million to $155 million. Non-GAAP operating margin should come in at roughly 24% to 26%. Given the total revenue and spending setup I described a moment ago, we anticipate Q3 non-GAAP EPS between $1.60 and $1.80; this EPS outlook presumes taxes of $55 million to $62 million, reflecting an estimated quarterly non-GAAP tax rate near 19%, and it further assumes a fully diluted share count of about 150 million shares. Turning to CapEx, for the reasons I mentioned before, we anticipate spending somewhere around $475 million to $525 million during the third quarter. That works out to roughly 43% to 46% of total revenue. As we look toward full year 2026, our revenue expectation is $4.445 billion to $4.53 billion, representing growth of 6% to 8% as reported and 5% to 7% in constant currency. Within Cloud infrastructure services, we still anticipate year-over-year revenue growth of no less than 50% in constant currency. We still anticipate Security revenue growth in the high single digits on a constant currency basis in 2026. And for delivery and other cloud apps, we still anticipate a year-over-year decline in the mid-single digits on a constant currency basis.

Edward McGowan

executive
#4

If we look at spot rates right now, our guidance bakes in a $9 million favorable foreign exchange effect on revenue for 2026 compared with the prior year. Because the U.S. dollar has strengthened lately, that number is much smaller than the positive $20 million figure we mentioned during last quarter's earnings call. Now on to operating margins: for 2026, we estimate a non-GAAP operating margin of roughly 25% to 26%, using today's FX rates. Shifting to CapEx, at this point we expect full year capital expenditures to come in at about 40% of total revenue. Onto EPS: for the full year 2026, we anticipate non-GAAP earnings per diluted share between $6.40 and $7.05. Those non-GAAP profit projections assume an effective tax rate on a non-GAAP basis of about 19%, plus a fully diluted share count near 150 million. One more thing before I wrap up: I want to add some detail about the way we safeguard profitability and control risk across our Cloud infrastructure services segment, especially when it comes to the sizable agreements we recently made public. Even though the money goes out the door at the start, these deals are structured so they generate robust cash flows across their lifespan, backed by take-or-pay obligations. So you can see how we approach such opportunities, let me walk through the economics that tend to show up in the multi-hundred million dollar, multiyear contracts we've encountered so far and anticipate going forward. I can't discuss any specific contract, but here's what I'd offer as an illustration of the big deals we've closed and what's showing up in our pipeline right now. When it comes to these multi-megawatt, large-scale rollouts, a number of things factor into our evaluation — the credit quality of the customer, how long the contract runs, which compute architecture gets deployed (that means GPU or CPU), the availability of data center space and power, CapEx, and direct operating expenses like networking and power, plus maintenance for both software and hardware. Once you weigh all of that together, what you see across our signed deals and the active pipeline is a consistently very attractive profile, with non-GAAP cash gross margins running anywhere from the mid-60s to the mid-70s. Once you account for depreciation on hardware plus other operating costs, these big contracts usually produce non-GAAP operating margins somewhere between the low-to-mid 20s and as high as the low 30s. That covers everything, so I'll close it out there, and Tom and I would be glad to answer whatever you'd like to ask. Operator?

Samik Chatterjee

analyst
#5

Congratulations on this latest win — a customer win. Perhaps you could help us think through what you're observing regarding greater customer interest in the CIS business when it comes to pipeline. You've announced these kinds of significant deals, but as you're pointing out, you're also sold out; so does that in some way prevent you from signing any additional significant award wins with customers soon as well; so any visibility on both those fronts would be helpful.

F. Leighton

executive
#6

Yes, a very strong pipeline really spanning industries and deal sizes; and no, that doesn't stop us from bringing on more customers because obviously we're ordering more hardware. And usually, we're looking at roughly a 6- to 9-month window on the larger side. So no, we're continuing to sign customers up. Ed, do you have more color on that?

Edward McGowan

executive
#7

Yeah, I was simply going to chime in. So here's how to think about that commentary: last quarter we gave you a bit of a preview that, given what the pipeline showed us, there was a solid possibility we'd put in an order for additional GPUs, and we managed to sell all of it. A portion of that big deal is going to consume some of the inventory that's still left. And we're putting in a further order with NVIDIA for a sizable batch of extra GPUs — part of it to fulfill the new order, though much of that is already covered by what we had on hand. The remainder is driven by the robust pipeline we're seeing for more GPU demand. Now, regarding how the market is moving right now, in many instances customers are basically preordering, since plenty of them don't keep much inventory available. So typically, the discussions revolve around securing some GPU capacity. By and large, customers are looking to lock that in over a lengthy stretch of time. So a decent comparison is to think of it like a reserved instance, which is what people do in the cloud space. And as a rule we'll book that months ahead of time. So it really lets us keep solid inventory, so to speak, without stretching ourselves too far into the future, while still satisfying what the market needs.

Jackson Ader

analyst
#8

My first question concerned the lag between when these multiyear contracts get signed — the multimillion-dollar multiyear ones — and when you anticipate they'll actually start flowing into revenue. Tom, you mentioned something along the lines of a 6- to 9-month window. What I wanted to know is whether you mean the period stretching from signing the deal all the way to revenue showing up, or does that window describe a different thing?

Edward McGowan

executive
#9

Sure, Tom, I'll handle that for you. Typically it covers the span from putting pen to paper up until the point revenue gets recognized. With the truly big contracts, we normally won't carry that kind of inventory — take, for instance, the deal we referenced during the prior quarter — we aren't purchasing that volume on a speculative basis. Instead, a major customer will come in and place an order, and that drives things. Usually in those situations, you might require extra data center capacity. We've got a solid pipeline of data center capacity due to come online from now until the close of next year, and honestly that's a core strength of ours regardless. So for us it amounts to a routine process. That said, regarding the gap between signing and revenue recognition, with these sizable contracts it's typically 6 to 9 months. And if they're big enough, we'll flag them during the call, just as we've done across the last 3 calls, and provide some sense of when we expect those contracts to begin producing revenue.

Jackson Ader

analyst
#10

Okay. Understood, that's clear. Now one more quick question. Ed, it was genuinely useful to get that kind of illustrative walkthrough of what the margin profile looks like on these deals. But has it stayed fairly steady while component pricing and your own GPU pricing have moved around? And could there be a risk that part of that margin profile shifts down in a structural way if components face more upward pressure down the road?

Edward McGowan

executive
#11

Yes. That's a great question. Let me put it in terms of a term — regarding the pricing environment with — particularly with the large customers. I would describe it as somewhat more collaborative. What I mean is that everyone generally understands inventory is tight, and data center availability is tight. And customers get that, first, if they don't place an order when they say they will, the goods might not be available and prices might rise. So in the marketplace, prices move fairly quickly, and any kind of pricing increase, we can pass it through. And typically, we'll build that into a big contract. So say, for instance, I'm taking delivery of goods across a span of 6 to 9 months or so. Should pricing move, whether that's memory or hardware costs or similar items, we do have ways to account for that. And pricing gets adjusted by us all the time. So for instance, someone might place a follow-on order after buying 6 months ago and now they're purchasing today, we're acquiring the next batch of inventory, and that comes at a higher price; that price simply got marked right up by us. But with customers it's very collaborative, and in CIS, price increases are something we're able to get passed along.

John DiFucci

analyst
#12

Hey, good work, everyone. I truly mean that — what we're seeing is quite striking; and Ed, I genuinely value the thoroughness you provided regarding certain attributes, the margin attributes tied to those agreements. The magnitude of them and the sales that will flow from it is quite striking, but the earnings are equally so; but may I probe a bit regarding — you offered some figures across the duration of those contracts. Yet simply in — based on what other firms handling comparable arrangements do, it appears that initially you'll likely witness reduced earnings generated by such things. You'll need to begin outlaying funds and racking up costs prior even to sales, I'd presume, and afterward they build up gradually. Therefore across the span of a lengthier arrangement, suppose, a 12-month contract or whatever the case may be, at what point do you reach that earnings level? Since I simply don't want folks to be caught off guard if initially, you're observing added strain on the earnings line until you arrive at where you must be.

Edward McGowan

executive
#13

Yeah; that's an excellent inquiry, John. And the response is it varies somewhat; it's additionally a dynamic present within the operation itself, correct? Thus frequently, you'll secure data center capacity in advance. So I'll discuss the major agreements, in which you may launch, for instance, 6 data centers or whatever the situation might be alongside a specific client and the hardware arrives across a span of several months or similar. Consequently you'll certainly face colocation expenses prior to sales, based upon how the agreement is drafted, yet broadly, once the hardware is ordered and we begin recording depreciation, there's a comparatively brief window from when you rack and stack and bring the hardware online to generating sales. Another factor to remember as well is the accounting standards governing both sales and for colocation or leases actually. When you've got escalators and such items, you'll typically distribute that across the whole agreement. Thus your expenses get amortized evenly, your sales get amortized evenly. So you reach that complete profitability fairly swiftly; yet there could be a quarter featuring a bit of disruption where certain expenses land earlier. Still, broadly, there's a fairly swift climb to attain that profitability considering what we've observed thus far, in the pipeline as well as with what we've signed concerning clients eager to bring up this capacity and more of, as I mentioned, a reserve instance type capacity where you grant them a set amount of up to capacity and they possess the entitlement to utilize it across the duration of the agreement. It's merely a matter of how swiftly you can bring it up. When you're building out your models, I'd anticipate somewhat reduced margins during the period we're scaling into these, and after that they ought to climb toward the figures I mentioned.

John DiFucci

analyst
#14

Alright, that's helpful. So, if you describe it as a fairly brief period — something like 3 months, perhaps? I realize each deal isn't the same, of course, but...

Edward McGowan

executive
#15

Correct, 3 months. At the extreme, it could stretch to 5 or 6 months, though typically it lands at roughly 3 months.

John DiFucci

analyst
#16

All right, that's great. And if I may, I want to piggyback on what Jackson asked, since I believe his line of questioning is exactly where my own thinking was headed. When it comes to certain deliveries scheduled further out — you brought up an arrangement, for instance, where capacity gets handed over across perhaps 4 data centers progressively — not everything is going to be up and running within the coming year. What I mean is, it arrives at a later point. And you touched on quite a bit regarding how everyone's attention is glued to inflation, specifically inflation in components. For some of these agreements — the ones with a longer horizon, where delivery stretches further out — do they follow a cost-plus structure, guaranteeing you a specific margin? And does that margin stay shielded even if inflation surges the way it has across the last year or 2?

Edward McGowan

executive
#17

Yes. What I'd say is, each arrangement differs somewhat. That said, in broad terms, let's begin with the procurement angle, then we'll move on to the revenue angle. On procurement, whenever we sign a contract for a data center, our typical approach is to lock in terms as fixed as possible, for the longest duration we can manage. Power carries a bit of variability. Usually, what happens is you commit to some form of initial escalator — a couple of percentage points, roughly. And as I mentioned, under the accounting rules, you wind up straight-lining it, the same way you'd handle rent when there's free rent or an escalator built into rent, that kind of situation. So that typically takes care of it; we have a pretty clear picture from the start regarding what power will cost us, and hardware, clearly, is locked. Usually, should there be any hardware maintenance involved, you'd work that out in advance and put a long-term agreement in place. Then you build the revenue contract to mirror that. So you might include an escalator year by year to address rising labor expenses, your power cost escalations, and similar items; yet on the whole, you're able to lock down a solid portion of it across time. Where lead times run long — suppose 20% of the capacity gets stood up during the first quarter, followed by 30%, then 50% in the final quarter, and say the whole ramp takes 9 months — if pricing shifts around, with memory for instance, you'll have anticipated that and typically built a mechanism into the contract to handle it; so as I said, it's highly collaborative with the customers; we hold these sorts of discussions and make sure we're as protected as we possibly can be. And broadly speaking, up to this point, things have played out the way we anticipated.

Param Singh

analyst
#18

First off, regarding compute, it's encouraging to see this level of visibility and pipeline. I just want to get a sense of when you consider available capacity, what do you have in megawatts? What does your pipeline look like? How far can you expand? And I'd also like to know, to support this growth, would you have to add more sites beyond those you currently have? And then I have one more question after that.

F. Leighton

executive
#19

Yeah. We keep growing the platform across additional locations, with GPUs in more cities, and I believe that trend will persist moving forward. And we're positioned very well to secure the data center capacity we require. There probably isn't anyone who works with more data center companies than we do. Our service currently operates in 700 cities globally, across 130 countries. We've earned an excellent reputation as a dependable partner, with solid financials and long-standing relationships. So we're well placed to obtain whatever we need to serve these extremely large customers.

Param Singh

analyst
#20

Then for my second question, it's encouraging to hear what you said about Agentic. What's your view on how quickly that area of the business could speed up, particularly when modules like API security and microsegmentation start to gain traction as Agentic AI traffic arrives on the platform?

F. Leighton

executive
#21

Yes, across many dimensions there's a strong synergy; we apply AI to make our offerings far more capable. Through API security, we assist our customers in spotting their shadow AI. With LayerX now, we assist them in safeguarding the workforce. So when employees use AI, they aren't accidentally exfiltrating corporate — sensitive corporate data; and naturally, within the compute business, we're turning on the AI on our platform; so we're well positioned to back our customers' AI usage while also securing it.

Rishi Jaluria

analyst
#22

It's good to see the CIS business keeping up its momentum, plus a few significant customer wins. I have two things. First, aside from those big CIS agreements you just mentioned, many of us keep discussing the chance for genuine AI inferencing at the edge, taking advantage of the edge network you've built over a long time — the biggest such network anywhere. Could you speak to what opportunities you see on that front? And might some of these sizable AI deals driven more by central cloud begin extending further into edge cloud, particularly since you're the sole provider offering both edge cloud and central cloud through a single platform? Perhaps shed some light on that for us, and then I have a brief follow-up.

F. Leighton

executive
#23

Yes. You put that quite nicely, I'd say. It truly is a continuum, and one major benefit on our side is the ability to handle the core for enterprise customers looking to train their model, while also handling the edge for inferencing in cases where latency or bandwidth constraints matter. Robotics is an excellent illustration, for instance. A robot carries sensors, possibly including something like video; it perceives the surroundings. Getting that video input and ingesting it requires high bandwidth. Our deployed GPUs happen to be extremely well matched to handling that video, then applying AI to determine what's happening, and afterward using AI to send an instruction back to the robot about what action to take; and in lots of these cases, latency matters too, not just the video's bandwidth, but quickly telling the robot what to do. Consider the robot as a car, a driverless vehicle, or robots in a setting alongside humans, or inside a factory where acting at the wrong moment could create a problem. So that's a strong illustration where, sure, perhaps training gets done most efficiently in a more centralized manner, yet inferencing and using it while operating — well, there are cases where you'd want that done very near the robot. We observe this throughout the whole spectrum, and having everything on a single platform gives us a significant edge. To be clear, training foundation models — those enormous models — isn't something we do. But using models, or training medium or smaller ones, yes, that performs really well on our platform; and what matters is placing the right resource in the right spot for whatever your agent or model aims to accomplish. That way you optimize cost and you optimize performance.

Rishi Jaluria

analyst
#24

Great, that's really useful. Now perhaps regarding security at this moment, particularly given that the LayerX acquisition is arriving once it finalizes. Yet as we witness something like the spread of agents, correct, it's — each brand needs to determine how to both adopt them and safeguard itself, especially against not only malicious ones, but even rival agents. Could you perhaps take us through how we ought to consider that possibly turning into a boost for the security unit and where LayerX could slot into that broader plan?

F. Leighton

executive
#25

Sure. Regarding LayerX, consider it as safeguarding your staff who, for entirely valid reasons, utilize various AI tools, agents, models, and similar things. The risk, naturally, is that they end up exposing confidential information while using those instruments; that is precisely what LayerX stops — it detects and stops it. Plus it applies the corporate policies that a security team has established for an organization. Now handling agents actually extends well past that. As a matter of fact, we currently lead with our bot management and agent management offerings since our clients, with their websites and applications, are seeing more and more agents arrive, possibly in place of humans. Certain agents are perfectly acceptable. A user has authorized them. Naturally, we want to confirm that is true and then provide them with excellent service. However, many agents are performing actions lacking authorization. We assist our client by determining which agent it is, whether it's authorized, and what it's doing. After that, our client informs us what action they'd like us to take in reply. The LLM search engines are a solid illustration. Naturally, it's extremely important these days that your brands appear in the upcoming generation of search engines. To assist with that, we detect that, hey, this is a scraper originating from a search engine, and we supply it with distinct content compared to the standard app or the site; and we provide it with content containing all the correct keywords and also arranged so that the agent can — the scraper can handle it far more effectively. Consequently, the client receives significantly improved search rankings; and that's merely a single instance involving an agent. That occurs to be one instance of a scraper, yet there is simply a multitude of various situations, and that is what we assist our clients in doing: determining what it is, whether it's legitimate, and then taking the suitable action for their business.

Sanjit Singh

analyst
#26

During the period — a year or under — since CIS was established as its own operation, you've generated close to $3 billion in bookings, which is really remarkable; based on what you've said about the pipeline, it appears these trend lines will keep going. So from the team I was hoping to understand how much, or what kind of, financing approaches lie ahead? Also, how far is the company prepared to move into a net debt position, or to tap additional equity capital, in order to fund what looks like a very robust pipeline of future business?

F. Leighton

executive
#27

Yes, I'll begin, and then pass it to Ed. To start, this business has been a focus of our work for years, and it rests on a platform that we've spent decades building. So it isn't something that happened overnight. Still, you're correct that really within the past year, we've increased our investment in go-to-market, because CIS and the platform — our cloud platform — is now at a point where we can serve major enterprise customers at scale. As a result, you're witnessing very rapid growth. And we're fortunate to hold more than $4 billion of cash, and we intend to invest so this platform keeps growing. And Ed, would you like to provide some further details on that?

Edward McGowan

executive
#28

Yeah, absolutely. And like we discussed, the balance sheet currently holds $4.6 billion in cash. Our debt has been set up so that roughly every 2 years there's a payment — or a ladder, so to speak, of debt retirements ahead; but even with these large transactions, once the initial capital is put to work, free cash flow is outstanding; so you go down somewhat at this point. And then when you begin recovering part of that on the flip side, free cash flow margins on these deals are extremely high; that being said, our banking partners are excellent. Working with so many incredible firms has been such a fortunate situation for us, and their advice is excellent. Up to now, converts have proven very appealing to us and provided the best economics. Alongside our Board, we constantly examine various strategies and scenarios and consider what serves our shareholders best. Should we need it, there remains substantial debt capacity. If doing so is the correct move, we'll do it. Thus far, equity-linked debt is what we've used — no equity offerings have been done. But whatever opportunity is best and makes the most sense, the lowest cost of capital for us, we'll consider it. Merely within the model we currently operate in, I believe we can expand the business substantially; an investment-grade credit rating is still something we hold, which matters and assists us with our colocation providers. However, you've heard us mention pausing the buyback for now so that all available cash goes toward growth, since an opportunity like this hasn't been seen by us in an extremely, extremely long time. Capturing that is something I want to ensure; yet regarding our capacity to finance future growth, I'm quite confident.

Sanjit Singh

analyst
#29

Got it. Now for my next question, when it comes to how CIS growth is trending, we've seen 2 quarters where growth was below 40%, a bit under the level we finished last year at, yet your target is at least 50% growth across the entire year. Regarding how you're turning those large contracts into something that can actually generate revenue, how should we frame the ramp from Q3 to Q4? Should we expect CIS growth to stay steady and then spike sharply? And why are you confident that sharp spike will show up in Q4?

Edward McGowan

executive
#30

Yes. That's a good question. We don't give quarterly guidance. What I can tell you is that we mentioned receiving some of our GPUs somewhat later than anticipated; those slipped out of Q2 and into Q3. That means a handful fewer weeks of revenue from deals that were signed and had been set to ramp during Q3; so for CIS in Q3, I certainly am not anticipating any acceleration in growth. In Q4, though, I do anticipate a significant spike upward, and there are several reasons for that. First, we can now begin collecting revenue on the GPUs we sold, which we brought into inventory, and start racking and stacking them while earning revenue on those contracts. For the large contracts, I've provided some guidance on when I believed revenue would begin flowing, and those do begin in Q4. As of now, when we review everything, it remains on track. This isn't about needing to execute or get anything signed. It's purely about the time required for everything to be delivered, installed and producing revenue; so at this point, we believe everything is on track, and we're expecting a significant spike upward. That ought to keep ramping quite substantially into Q1 of next year too, since you capture a partial quarter of revenue from some fairly large deals and then far more begins flowing as you move into Q1. And I anticipate that acceleration will persist through the whole of next year.

Frank Louthan

analyst
#31

When you're doing CIS, how hard is it to secure power at those facilities? Also, could you update us on your current number of locations? And for this new business you just signed, how many will need to be built out?

F. Leighton

executive
#32

Yes. That's a great question. Securing the data center capacity and power we require puts us in a truly strong spot. This industry is one we've occupied for a very long time. It's likely that we work with a greater number of data center companies than anyone else. Right now, our infrastructure is deployed across more than 700 cities, spanning 130 countries. Through many years, we've earned an excellent reputation as a dependable business partner, thanks to financial strength and a very robust business. That means we can obtain the capacity we need. Our data center footprint keeps expanding, with additional locations; and as we grow, we're taking on bigger locations and more power. I'd anticipate that trend continuing, and our position there is really solid.

Edward McGowan

executive
#33

Right, and there's something else I'd like to mention. Frank, one of the benefits on our side is how remarkably adaptable we are, since we operate such a vast backbone. Take getting power on the West Coast — we've got a lot of choices about where to locate. Data center providers find that quite appealing, because power quality isn't uniform, even from one state to the next. So you could move a couple hundred miles out of a very costly state and land in one that's less expensive, serve as an anchor tenant for a developer putting up data centers, sign a long-term deal, secure committed power with strong performance tied into our backbone, and so on — and that lets you meet whatever needs that specific customer has while achieving far better economics, since the power costs less. So that's another edge we hold: we have tremendous, tremendous adaptability regarding where we can expand. And we're appealing enough to these major data center builders because we aren't requesting a gigawatt facility. Often what we're discussing is 5 to 10, perhaps 15 or 20 megawatts — a commitment large enough to genuinely assist them in scaling, earning a solid IRR on their investment as they proceed, and obtaining credit to construct those facilities.

F. Leighton

executive
#34

Correct. Ed makes an excellent observation there. It isn't something most folks spend much time considering, yet we run one of the biggest backbones on the planet. Naturally, we rely on that backbone to link up every one of our sites. Selling connectivity as such isn't what we do. That distinction matters a great deal whenever we move into fresh markets. We're able to enter areas quite effectively and hook them into our platform in cases where competitors might struggle.

Rudy Kessinger

analyst
#35

I realize you don't typically discuss these huge agreements or the applications involved much. But if you can, perhaps address whether we're talking about this robotics contract, a few of those other big ones, or simply what's sitting in the pipeline — which particular measurements, be it latency or something similar, are those clients evaluating when they conclude that Akamai offers a clear advantage over the NeoClouds and the hyperscalers, and that's why we'll take this path with Akamai? And which concrete performance benefits are causing you to be selected for certain of these sizable workloads?

F. Leighton

executive
#36

Yes. That's an excellent question. We're unable to discuss the particular agreements. Broadly speaking, though, Akamai gets picked owing to a stellar reputation we've earned the hard way when it comes to dependability. Customers are able to put their faith in us. Ours is an enormously distributed platform that stands out as genuinely distinctive within the market — as I noted, 700 cities across 130 countries; we're able to place our clients' compute requirements near their users, near their data, which yields superior performance, reduced latency, improved scalability, especially when video is part of the picture. We offer full stack compute alongside storage, paired with the top cybersecurity offerings globally, the top delivery platform globally, and at an excellent price, since having the top delivery platform globally plus this enormously distributed capacity means our egress expenses run far below what rivals charge. Consequently, we're able to present an attractive value proposition: excellent performance, excellent dependability, at reduced expense.

Fatima Boolani

analyst
#37

Ed, I'd like to discuss the additional capacity you're adding, which you described using a reserved instance framework. That approach provides considerable visibility for you and a great deal of predictability for customers. However, I'm aware you've considered establishing a rental offering inside the boundaries of your CIS operation. So I'd like to understand where things stand in that effort, and if you happen to possess spare capacity that could support a more spot-oriented pricing approach, perhaps leading to quicker expansion in CIS, and whether that matters for how we should think about this year. Then I have another question about the security business, please.

Edward McGowan

executive
#38

Right; Fatima, we provide both options. What we're seeing, though, is that demand is stronger among customers who prefer to commit for an extended duration. Those arrangements are clearly more favorable for us since they let us secure value across a longer stretch, and you receive a guaranteed ROI on whatever you're putting money into. Both are available from us. Some clients even approach us requesting alternative structures — basically saying that if we have spare capacity, they'll take it for some timeframe and just want advance notice should we need it returned. So the market, I'd say, tends to lean — at least among our own customers — toward those seeking certainty rather than simply renting on an hourly basis, though we do provide both. And I'd say the bulk of current growth, clearly, because of how large these deals are, comes from the more committed approach, but both remain available.

Fatima Boolani

analyst
#39

I appreciate that. And regarding the security business, you pointed out the tailwinds you're experiencing specifically in Guardicore Segmentation, API security, driven by how traffic is mixed on the Internet and, as a result, your platform evolving. I'm curious whether you could address some of your more traditional and perhaps more mature product areas, such as DDoS and, I'd say, classic WAFs. What's happening there? Is a rising tide lifting all boats, and is there perhaps a renaissance in customer spending, maybe a renewed focus? I'd love to gain a clearer picture of how those traditional and more mature parts of the security product portfolio are performing compared with some of your faster-growing SKUs in the platform.

F. Leighton

executive
#40

Great question. Yes. And we have witnessed genuine tailwinds from AI and, in a sense, the post-Mythos era for our longer-term security offerings. I think it's roughly because of AI that attackers have put together far bigger bot armies to carry out attacks. And the scale of those attacks has risen by perhaps a factor of 10 during the last year; and so our DDoS services are more needed. And with web app firewall, particularly there, since now there will be far more zero days. We detect those before anyone else does, and our firewall policies are refreshed so our customers are protected ahead of a zero-day becoming widely known. That gives them the opportunity to complete any patching they require, and to do it without risk. I've had conversations with a great number of CISOs, CIOs, and CEOs in recent months, and a lot of them have urgent initiatives in motion to ensure every one of their applications and sites sits behind our web app firewall, since they understand that if Akamai's web app firewall is in front of it, they'll be fine whenever zero-days inevitably appear. AI is providing strong momentum everywhere we look, and this creates an opening for Akamai to genuinely assist its largest enterprise accounts.

Konark Gupta

analyst
#41

Konark here, filling in for Patrick. I wanted to ask whether you might elaborate on that CrowdStrike customer win — a genuinely encouraging indicator, in my view, for the sort of architecture Akamai has. Could you share a bit about the timing of when that deal was finalized, and whether any partnership elements exist alongside the customer relationship?

F. Leighton

executive
#42

Sure, it was finalized not long ago. It serves as a genuine proof point for what our security offerings can do. Clearly, CrowdStrike is a firm that places enormous importance on both security and reliability. They had — as they've mentioned publicly — dissatisfaction with the provider they had before, and what drew them to Akamai was our reliability along with the stronger security capabilities we bring them; and we do have a partnership with them. So that makes for a solid relationship too. Still, I view it as a strong endorsement of Akamai and what we deliver in security.

Operator

operator
#43

Alright folks, that brings us to the end of both the Q&A portion and today's call overall. We appreciate everyone taking part and being with us today. At this point, feel free to hang up your lines.

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.