Akamai Technologies Inc

NASDAQ:AKAM

Earnings Call Transcript 2022-Q3

[Operator]

Good day, and welcome to the Akamai Technologies Third Quarter 2022 Earnings Conference Call. [Operator Instructions].

I would now like to turn the conference over to Tom Barth, Head of Investor Relations. Please go ahead, sir.

[Tom Barth, Head, IR]

Thank you, operator. Good afternoon, everyone, and thank you for joining Akamai's Third Quarter 2022 Earnings Call. Speaking today will be Tom Leighton, Akamai's Chief Executive Officer; and Ed McGowan, Akamai's Chief Financial Officer.

Please note that today's comments include forward-looking statements, including statements regarding revenue and earnings guidance. These forward-looking statements are subject to risks and uncertainties and involve a number of factors that could cause actual results to differ materially from those expressed or implied by such statements. The factors include any impact from macroeconomic trends, the integration of any acquisitions and any impact from geopolitical developments.

Additional information concerning these factors is contained in Akamai's filings with the SEC, including our annual report on Form 10-K and quarterly reports on Form 10-Q. The forward-looking statements included in this call represent the company's view on November 8, 2022. Akamai disclaims any obligation to update these statements to reflect new information, future events or circumstances, except as required by law.

As a reminder, we'll be referring to some non-GAAP financial metrics during today's call. A detailed reconciliation of GAAP and non-GAAP metrics can be found under the financial portion of the Investor Relations section of akamai.com.

And with that, let me turn the call over to Tom.

[Thomson Leighton, Co-Founder, CEO & Director]

Thanks, Tom, and thank you all for joining us today. I'm pleased to report that Akamai delivered strong results in the third quarter despite the ongoing challenges with the global economic environment and the effects of a strong U.S. dollar.

Q3 revenue was $882 million, up 3% year-over-year and up 7% in constant currency. This result was driven by the continued strong growth of our security and compute businesses, which collectively grew 23% year-over-year and 28% in constant currency. These 2 business lines accounted for 55% of our overall revenue in the quarter.

Q3 non-GAAP operating margin was 28%. And non-GAAP EPS was $1.26 per diluted share, down 13% year-over-year or down 7% in constant currency. EPS was negatively impacted once again by foreign exchange rates and a higher effective tax rate compared to last year. Free cash flow was very strong at $271 million in Q3, and it amounted to 31% of our revenue.

I'll now say a few words about each of our 3 main lines of business, starting with security. Our Security solutions generated revenue of $380 million in Q3, up 13% year-over-year and up 19% in constant currency. The growth was particularly strong for our enterprise Zero Trust products, which were up 51% year-over-year in constant currency.

Our Guardicore Segmentation solution continued to lead the way with several major customer wins. For example, one of the largest energy companies in the world adopted Guardicore to help protect against SolarWinds types of ransomware attacks. A leading global developer of dietary supplements adopted our Segmentation solution to help meet European regulations and limit cybersecurity risk. And a major South American broadcaster deployed Guardicore to protect their reporting of election results.

Our market-leading app and API protection products also performed well in Q3, with many wins against the competition. For example, the largest bank in Southeast Asia came to Akamai last quarter after suffering repeated outages by a competitor that had lured them in with low pricing. When the bank faced large fines from regulators for the extended outages, they saw more value in being back on Akamai's platform.

One of the top banks in North America is in the process of bringing all of their traffic back to Akamai after struggling with outages at another competitor who would also lure them in with lower pricing, but couldn't deliver the performance and reliability needed by a major enterprise. After testing our capabilities against competitors, one of the world's largest financial services companies expanded their relationship with us, contracting for 10 of our products and services, including Bot Manager and Page Integrity Manager.

A Fortune 100 food processor and commodities trader became a new Akamai customer last quarter after an anonymous threat drove them to seek better DDoS and web app protection than they were getting from a competitor. And in Germany, an online advertising business with one of the country's busiest websites suffered severe account takeover attacks, load problems and reputation damage before coming to Akamai for bot management and app and API protection.

Given such examples, it's not surprising that Akamai's web app and API protection was named a leader in both Gartner's Magic Quadrant and in Forrester's Wave report last quarter. Our compute product group also performed well in Q3 with revenue of $109 million, up 72% year-over-year and up 77% in constant currency.

We're continuing to make good progress on integrating Linode into our edge platform and on adding the capabilities and scale needed to support mission-critical applications for major enterprises. In particular, we've connected all of Linode's 11 existing locations into our private backbone, enabling us to provide lower latency, higher throughput and improved egress economics. We've also expanded the capacity of these facilities and are in the process of adding 13 additional sites, 5 of which are expected to go live in Q1 with 8 more planned for Q2.

As we discussed at our Analyst Day in May, we're also developing a lighter-weight deployment model that is suitable for distribution at a broad scale. This will enable us to get compute much closer to end users around the world. We plan to deploy several dozen of these lighter-weight sites next year, at which point we expect to compare well with the hyperscalers in terms of points of presence and proximity to both enterprise data centers and end users.

Of course, we plan to have all of our compute sites integrated into Akamai's unique edge platform, which has over 4,000 locations for edge computing. As a result, we expect to be able to offer superior performance as well as lower total cost of ownership for enterprise computing needs.

We've also made significant progress on adding new and improved enterprise capabilities to our compute platform. We launched Database-as-a-Service with managed MySQL in May and managed Postgres in June. We released the next generation of our Kubernetes platform in Q3 to enhance performance and reliability. We expect to launch early versions of an enhanced object storage product as well as next-gen serverless capabilities next quarter. And we expect to become SOC 2 and ISO 27001 compliant this quarter with PCI compliance expected to follow in the first half of 2023.

Although we still have much work to do, we're encouraged by the customer use cases that our compute platform began serving in Q3. One of the world's top development studios for gaming moved their matchmaking service to Akamai to help them with data processing and analysis. A large online legal services platform in India chose Akamai as part of their multi-cloud strategy after they concluded that we could help them optimize their cloud computing budget.

And a large media workflow company in Germany is planning to migrate their apps from a hyperscaler to Akamai, calling our new capabilities a great addition, especially with the plans for a high number of distributed sites and the tight integration with Akamai content delivery.

Over the past few months, I've spoken with many of the world's leading enterprises about our plans for cloud computing. Most tell me that they want more choice in cloud computing, and they often express concern about being locked into contracts with cloud giants that are consuming larger portions of their IT budgets, especially in cases when their cloud vendor is also a direct competitor. Customers also understand the value of leveraging a more widely distributed cloud platform and one that directly connects to Akamai's unique edge platform with over 4,000 points of presence.

Turning now to our CDN business. Our delivery products generated revenue of $393 million in Q3, down 15% year-over-year and down 11% in constant currency. These results reflect continued deceleration in traffic growth among our largest customers and the impact of some large renewals that we completed in the first half of the year.

As we said at our Analyst Day in May, we've aligned our pricing strategy with the slower traffic growth rates we've experienced this year. In addition to scaling back discounts upon renewal, we're continuing to decline business from a very small number of customers who have extreme traffic peaks compared to their daily usage patterns.

While this resulted in less revenue in Q3, it's enabled us to meaningfully lower our delivery network CapEx as we direct cash flow from our delivery business to our compute and security businesses where we have a higher ROI.

As Ed will detail shortly, we're also taking several steps to reduce OpEx, including reducing our real estate footprint and limiting hiring to our most critical areas. Although we're facing the same challenging macroeconomic environment as other companies, I believe that Akamai is on the right path to long-term growth and success with our disciplined management of expenses and strong focus on opportunities for future growth such as cloud computing.

Becoming a force in the enormous cloud computing market won't be easy, but I believe that it's something that Akamai can accomplish. Akamai has a strong track record of continuous innovation and business expansion. Along the way, we've achieved significant milestones that many thought were impossible.

In our first decade, we pioneered the CDN industry. a multibillion dollar market, where we remain the leader by far. In our second decade, we created the industry for app and API protection as a cloud service, our second multibillion dollar market, where we are the leader by a wide margin. Looking ahead, Akamai is on the cusp of another major phase of expansion with our foray into cloud computing.

Having already scaled content delivery and cloud security into billion dollar businesses, we now have an opportunity to do it again with cloud computing. In fact, I believe our opportunity in cloud computing is even larger than it's been for delivery and security.

Cloud computing is a $100 billion market growing at a very rapid rate. We believe we're in an excellent position to capture a share of this business, particularly from companies that value our market-leading delivery and security solutions and that don't want to be locked in to more expensive options with a cloud giant that competes against them.

Akamai is a company that enterprises can trust to be their partner, to scale with their business and to provide the best when it comes to security, reliability and performance. By adding compute to our unique edge platform, we can provide a full suite of cloud services that will help lower our customers' cost to build, run, deliver and secure their applications.

In summary, my confidence in Akamai's future prospects for growth and success has never been higher. In fact, my confidence in what I see ahead for Akamai has led me to take steps to put in place a 10b5-1 trading plan, not to sell, but to buy $3 million in Akamai stock over the next 6 months. We expect to announce the adoption of my plan in a formal filing later this week.

Now I'll turn the call over to Ed for more on Q3 and our outlook. Ed?

[Edward McGowan, CFO]

Thank you, Tom. As Tom mentioned, Akamai delivered a solid quarter in Q3 despite a very challenging macroeconomic environment. Q3 revenue was $882 million, up 3% year-over-year or 7% in constant currency. The stronger U.S. dollar negatively impacted our year-over-year growth rate by approximately 4 points or about $39 million of revenue year-over-year and $14 million on a sequential basis.

On a combined basis, our security and compute businesses represented 55% of total revenue, up 23% year-over-year and 28% in constant currency. Security revenue was $380 million and grew 13% year-over-year and 19% in constant currency, led by another strong contribution from Guardicore. Guardicore delivered approximately $14 million of revenue in Q3. Security represented 43% of total revenue in Q3, which is up 4 points from Q3 a year ago.

Compute revenue was $109 million in Q3, up 72% year-over-year and 77% in constant currency. As Tom mentioned, while we are in the early innings of our cloud computing journey, we are very excited about initial feedback from customers and the significant growth opportunity ahead.

Delivery revenue was $393 million, down 15% year-over-year and down 11% in constant currency. Sales in our international markets were $421 million and represented 48% of total revenue in Q3, up 1 point from Q2. International revenue was up 2% year-over-year or 12% in constant currency. Finally, revenue from our U.S. market was $461 million, up 3% year-over-year.

Moving now to costs and profitability. Cash gross margin was 75%. GAAP gross margin, which includes both depreciation and stock-based compensation, was 61%. Non-GAAP cash operating expenses were $291 million. Adjusted EBITDA was $368 million, and our adjusted EBITDA margin was 42%. Non-GAAP operating income was $243 million, and our non-GAAP operating margin was 28%. It is worth noting that on a year-over-year basis, our non-GAAP operating margin was negatively impacted by approximately 1 point due to unfavorable foreign exchange rates.

Capital expenditures in Q3, excluding equity compensation and capitalized interest expense, were $111 million. As we mentioned on our Q2 earnings call, our strategy in our delivery business is to be more selective on the peak traffic levels we will take on our network. As a result, delivery network CapEx, excluding Linode, was just under 4% of revenue in Q3.

GAAP net income for the third quarter was $108 million or $0.68 of earnings per diluted share. Non-GAAP net income was $200 million or $1.26 of earnings per diluted share, down 13% year-over-year and down 7% in constant currency. It's worth noting that on a year-over-year basis, foreign exchange rates negatively impacted our non-GAAP EPS by approximately $0.10 in Q3.

Taxes included in our non-GAAP earnings were $41 million based on a Q3 effective tax rate of approximately 17%. This was about 1 point higher than our guidance due to a more unfavorable mix between U.S. and foreign earnings.

Now moving to cash and our use of capital. As of September 30, our cash, cash equivalents and marketable securities totaled approximately $1.4 billion. During the third quarter, we spent approximately $163 million to repurchase shares, buying back approximately 1.8 million shares.

Our ongoing share repurchase activity has resulted in a net reduction in our non-GAAP fully diluted shares outstanding of approximately 5 million shares or roughly 3% on a year-over-year basis. We ended Q3 with approximately $1.4 billion remaining on our current repurchase authorization. Our intention is to continue to buy back shares to offset dilution from employee equity programs over time and to be opportunistic in both M&A and share repurchases.

Before I provide our Q4 outlook and an update to our 2022 guidance, I want to highlight several factors. First, with nearly half of our revenue coming from outside the U.S., the strong U.S. dollar continues to be a significant headwind to our reported results. At current spot rates, our guidance now assumes foreign exchange will have a negative $130 million impact to revenue in 2022 on a year-over-year basis.

As I mentioned previously, the strong dollar also impacts our margins and earnings. We estimate FX will negatively impact our non-GAAP operating margin by approximately 1 point year-over-year and non-GAAP earnings by approximately $0.34 for the full year 2022. Second, we have seen a lengthening in some of our sales cycles. We believe that is -- this primarily reflects the uncertain macroeconomic conditions that our customers are experiencing, and it is visible in many parts of our business.

Finally, we continue to closely monitor our costs in light of ongoing inflationary and macroeconomic pressures across the globe. We have made good initial progress on our cost-cutting measures that we mentioned on our last call, which include real estate costs, where we sublease some of our underutilized office space in Q3, and we'll continue to look for additional savings going forward.

Reducing our third-party cloud expense in 2023, where we look forward to making significant progress on shifting workloads to Linode. And lowering network CapEx associated with our delivery business, where I noted our continued progress on reducing spend significantly related to traffic delivery.

In addition to these items, as Tom mentioned, we plan to be very disciplined with headcount and focus our investments on higher growth areas like cloud computing and security. In particular, we are closing over 500 open positions and re-tasking many other employees to work on compute. These closures went into effect today.

And just a quick reminder about our typical fourth quarter dynamics before I turn to our Q4 guidance. As in prior years, seasonality plays a large role in determining our fourth quarter financial performance. We typically see higher-than-normal traffic for our large media customers and from seasonal online retail activity from our e-commerce customers, which are both difficult to predict, especially during this more challenging macroeconomic environment.

With that in mind, we are projecting Q4 revenue in the range of $890 million to $915 million or down 2% to up 1% as reported or up 3% to 6% in constant currency over Q4 2021. Foreign exchange fluctuations are expected to have a negative $11 million impact on Q4 revenue compared to Q3 levels and a negative $44 million impact year-over-year. At these revenue levels, we expect cash gross margins of approximately 74%. This roughly 1 point sequential decline is primarily driven by increased third-party cloud costs and some compute-related data center build-out costs.

Q4 non-GAAP operating expenses are projected to be $298 million to $306 million. We anticipate Q4 EBITDA margins of approximately 40% to 41%. We expect non-GAAP depreciation expense to be between $125 million to $126 million, and we expect non-GAAP operating margin to be approximately 27% for Q4.

Moving on to CapEx. We expect to spend approximately $122 million to $127 million, excluding equity compensation and capitalized interest in the fourth quarter. This represents approximately 14% of projected total revenue. And with the overall revenue and spend configuration I just outlined, we expect Q4 non-GAAP EPS in the range of $1.23 to $1.30. This EPS guidance assumes taxes of $38 million to $40 million based on an estimated quarterly non-GAAP tax rate of approximately 16%. It also reflects a fully diluted share count of approximately 158 million shares.

And finally, for the full year 2022, we now expect revenue of $3.58 billion to $3.6 billion, which is up 3% to 4% year-over-year as reported or up 7% to 8% in constant currency. We continue to expect security growth of approximately 20% in constant currency for the full year 2022.

We now estimate non-GAAP operating margin to be approximately 28% and non-GAAP earnings per diluted share of $5.23 to $5.30. And this non-GAAP earnings guidance is based on a non-GAAP effective tax rate of approximately 16.5%, a fully diluted share count of approximately 116 million shares. Finally, full year CapEx is anticipated to be approximately 13% of revenue.

In closing, we are very pleased with how the business is continuing to perform despite a very challenging macroeconomic backdrop. We are very excited about our future growth opportunities ahead. Thank you. Tom and I would be happy to take your questions. Operator?

[Operator]

[Operator Instructions]. Today's first question comes from Keith Weiss at Morgan Stanley.

[Keith Weiss, Morgan Stanley]

Excellent. Nice quarter in a difficult environment. On that difficult environment point, I was hoping you could help with a little bit more specificity in terms of where you guys are seeing the macro impacts. It sounds like security is still holding up relatively well. And on the delivery side of the equation, there's some Akamai-specific impacts there. Could you give us some kind of detail in terms of where the risk factors are, sort of where the macro lies on a product and geographic perspective? I think that would be helpful.

And then I guess on the CapEx side of the equation, you talked a little bit about types of business that you guys are not looking to take onboard on a go-forward basis, the very peaky workloads that perhaps were overly taxed in the system, if you will. Can that lead to a fundamental different kind of CapEx intensity for the business on a go-forward basis? Or is it just too small to make a difference?

[Edward McGowan, CFO]

Keith, this is Ed. I'll take those. I'll start with the second question first. So from a CapEx perspective, the way to think about it is in the delivery business, we had -- if you go back to our May Analyst Day, we talked about CapEx and delivery being in sort of the high single digits. We've been running in the lower single digits. We're just under 4%. So certainly, in the near term, it will have an impact in the delivery business. As the compute business gets larger, obviously, that's going to be the main driver for CapEx.

But certainly, you saw that in Q3, the impact of overall CapEx down at around 13%, and for the year, it's around 13%. So it is less capital-intensive. But as we look at our cloud compute business, obviously, there's going to be some -- Tom talked about building out in more locations. So there'll be some more CapEx associated with that. But we are seeing a pretty healthy decline in our delivery business, which is as anticipated.

On your first question, you asked about the macroeconomic environment where we're seeing some challenges. I mentioned in my prepared remarks that we are seeing some of our sales cycles lengthening, seeing customers pushing off upgrades for certain products and things like that, which is pretty typical. Most companies are doing that. I know we're -- we're doing something similar as we go through our budget cycles. We think it's temporary in nature.

Also seeing a little bit of pressure in some of the advertising-related businesses, a little bit of pressure there. And then Europe is something that we're keeping a pretty close eye on, that's about geographically where we're worried. Obviously, with what's going on with energy costs and things like that in Europe, that's something that we're keeping a very close eye on and have been a little bit cautious on our guidance, I would say, in Q4, just relative to seasonality. Just trying to take that into consideration.

[Keith Weiss, Morgan Stanley]

Got it. And on the energy cost, is that a top line concern in terms of you're worried about your -- what your customers are experiencing? Or is that more of a gross margin concern that you guys are worried that those energy costs can impact your gross margins on a go-forward basis?

[Edward McGowan, CFO]

Yes, I'd say it's more on our customers and their customers, so how does the consumer behave. Obviously, retail is a driver of seasonality as is spending for media. So those 2 things could be impacted. And then obviously, with our customers, if you see shutdown in manufacturing and things like that, that's obviously going to have a ripple effect on GDP across Europe. So that's from that perspective.

As far as our risk with energy, we do a pretty good job. The team has done a nice job with our colo negotiations. If you think about our costs, our server costs were pretty insulated from costs there on the bandwidth side that tends to be deflationary. Colo, there is some energy exposure, but the team has done a really good job of trying to lock in longer-term deals. We're not seeing that. It's possible that may start to affect us later into next year. But right now, we've got it pretty well under control.

[Operator]

And our next question today comes from James Breen at William Blair.

[James Breen, William Blair & Company]

Can you just talk a little bit about the compute business? I recognize it was up a lot year-over-year after you closed Linode. It was up a few million quarter-to-quarter. What do you have to do to accelerate that business? Is it building out more resources? Is it just you're getting some larger customers? And sort of what are the thoughts there on what that could ultimately grow? It seems like with the opportunity, it could grow faster than the security business.

[Thomson Leighton, Co-Founder, CEO & Director]

Yes. Great question. The compute business, even before Linode, was on a pretty strong trajectory of growth. And with Linode, that accelerates it a lot. As you look to the future, the big growth comes when we're tapping into the core cloud compute market, a market that's over $100 billion today and growing rapidly. And that's something that we're working really hard on now to -- so we can exploit that next year. And that involves increasing scale, having a lot more core compute regions and then introducing the lighter-weight distributed computing regions.

So that we'll be in a position to offer at least as good or better performance integration into the Akamai platform, which has great delivery, great security and, of course, edge computing at a lower total cost of ownership. And for a lot of our customers, especially you think of the media vertical and the commerce vertical, they spend a lot more on compute than they do with delivery and security. Moreover, they compete pretty heavily with the hyperscalers.

So the big growth for us, what we're really going after over the next several years is in that core cloud compute market, mission-critical applications for major enterprises because that's a very big potential market for us. And that's what will drive our -- the major growth in compute, and ultimately, I think the company going forward.

[Operator]

And our next question today comes from James Fish at Piper Sandler.

[James Fish, Piper Sandler]

I appreciate the questions. Obviously, I agree with you on deceleration in traffic overall, especially on the media side. But what makes you guys confident that you aren't losing traffic share of some of the media customers, especially as you're purposely not doing some of these large events, for example, or kind of the gaming peak traffic? And any sense to how much that's kind of impacting the media business this quarter in this year overall that we should kind of normalize as we start to think about for next year?

[Edward McGowan, CFO]

Jim, this is Ed. Good question. So actually with traffic, we are starting to see a little bit of a recovery in September. It continued a bit in October as well. But in general, it is, as we talked about, a much lower year relative to what we typically see.

As far as the specific customers, you're talking about only a handful of big customers that can drive this kind of peak. And these particular customers all have multi-CDN. So in this particular case, we did lose a couple of million dollars. It's not significant in terms of the impact on the year. But as you can see, it saved us about, call it, 4 points on CapEx. So it's pretty meaningful from an overall economic standpoint.

Now that said, we still -- these still are big -- very big customers of ours. They've just sort of flattened out the peak a bit. So as their daily average traffic is not growing as quickly, the economics just don't work for us anymore. So we just -- they're still good customers of ours, but just decided we weren't going to allow them to peak as much as they did in the past. It just makes sense for us. But it's not an overly material number. A few million dollars is the way to think about it.

[James Fish, Piper Sandler]

That's helpful. I appreciate that. Maybe following up a little bit on Keith's prior question around more specifically on the security growth, which slowed to about 15% when I normalize everything. What makes you guys confident that we're going to see an acceleration of this business back to that 20% all-in constant currency growth rate over the next couple of years? And is the impact today being more felt on the new business side, given kind of your comments along elongating sales cycles? Or is it you're seeing a slowdown in existing customers' expansion as well?

[Thomson Leighton, Co-Founder, CEO & Director]

Yes. Good question. The 20% goal, of course, includes M&A, and we're about to do the year-over-year lapping on Guardicore, which is a very successful acquisition. And we haven't announced any other acquisitions in security to sort of fill that gap. In terms of increasing the security growth rate over time, obviously, the global economic conditions are important there.

Also, we've got several of the newer products that are growing very rapidly. Aside from Guardicore, bot management doing extremely well. The new Account Protector solution doing very well. Page integrity management, which will be, I think, really important for companies that want to be PCI compliant beginning in 2025, and they're already working towards that.

Those areas are doing very well in terms of growth, but they're still small enough in revenue that they can't swing the whole number as much. The vast majority of our security revenue today is in the app and the API protection area. And the large majority of that is in our web app firewall, where we're the market leader by far and continuing to grow that faster than the market and our competitors there.

But that market is -- as a whole, is slower growing. So what we'll need to see is better economic environment, continued growth in the rapidly growing products that are -- as they get bigger, their rapid growth can drive security as a whole, and ultimately, M&A, which is a key part of the 20% goal.

[Operator]

And ladies and gentlemen, our next question today comes from Frank Louthan at Raymond James.

[Frank Louthan, Raymond James & Associates]

Great. With the sales cycle more elongated, can you give us a little more detail there? Is that across the board? Is it concentrated in any verticals? And give us a little bit more color on what's driving that. Is it more economic? Or is there anything to do with the mix with Linode and compute that's sort of making the suite of services take a little longer for folks to make a decision?

[Edward McGowan, CFO]

Frank, good question. So actually, I'll start with Linode. Actually, Linode in an environment like this, given what Tom talked about, we will be able to provide comparable services at a much better set of economics and actually think is an opportunity for us. So I would expect that as we go into next year, that should be a tailwind for us.

On the headwind side, we are seeing across the board, certainly from a geographic perspective, that sales cycles are elongating. We've seen some deals push. Probably the easiest place to identify it is in Guardicore. Guardicore, as Tom mentioned, is still doing really well, but we have a dedicated sales team. So it's a little bit easier to track those deals as they're going through the pipeline. Those deals also sometimes tend to be a little bit larger. So it's a little bit easier to follow that.

With our business, given it's a SaaS business and we've got recurring revenue contracts are constantly going through and renewing contracts, what you're seeing is some customers that are talking about, say, adding Page Integrity Manager or a Bot Manager or Account Protector, just pushing that off into the future as they go through their budget cycle.

So we're not seeing as much on the renewal side from an upgrade perspective. So a little bit of slowness there. And then from the new customer acquisition perspective, I think pretty much every tech company you talk to these days is seeing it a little bit harder to attract new customers. So it's a little bit of a combination of everything as we've just been impacted by this economic impact here.

[Operator]

And our next question today comes from Fatima Boolani with Citigroup.

[Unidentified Analyst]

This is on for Fatima. So just maybe a follow-up in regards to the September and October recovery on the delivery side you're starting to see. Can you maybe give a sense of which end market cohort is really driving that momentum? And then when you're going to negotiating tables in regards to pricing and other contract terms, can you also give us a sense of how that has been developing?

[Edward McGowan, CFO]

Sure. Yes. So in terms of the traffic sort of getting a little bit healthier, I would say, coming out of the summer months, media is probably the vertical that it's most obvious in. So we're starting to see that across video, a little bit in gaming. Gaming is still overall very, very weak compared to what it's been in the future, but a little bit of an uptick here in the gaming vertical over the last couple of months. And then your other question, can you just remind me again? I forget.

[Unidentified Analyst]

Sorry, just in terms of pricing and other contract terms as you guys go to the negotiation table.

[Edward McGowan, CFO]

Yes. So one of the things we talked about is, obviously, like in the media division in particular, media verticals in particular, as we see traffic levels decline or not grow as quickly, I should say, we typically would give a discount commensurate with what you see in the traffic growth rate.

So obviously, as traffic growth rates are not growing as quickly, we are lowering our discounts that we're providing to our customers, and we're starting to see that make its way through the system. It will take a while for it to really impact the growth as we go through our renewal cycles, but I am seeing that the pricing declines are certainly moderating a bit.

[Unidentified Analyst]

Okay. Got it. And then maybe just a follow-on, on the lighter-weight development. Any sense of how the customers -- are there any #1 customers in the beta testing phase? And also, I guess, where -- what are sort of the milestones we should look out there -- look out for there?

[Thomson Leighton, Co-Founder, CEO & Director]

I didn't catch the question. Can you repeat the question, please?

[Unidentified Analyst]

Sorry, just in terms of the lighter-weight deployment on, I guess, the cloud side? Yes.

[Thomson Leighton, Co-Founder, CEO & Director]

Yes. And what's the question about the lighter-weight deployment?

[Unidentified Analyst]

Yes. Just in terms of, are there any customers currently in the beta phase or testing out the product and how has that feedback been? And then are there any milestones we should look out for?

[Thomson Leighton, Co-Founder, CEO & Director]

Good. Okay. So there are customers on the platform today and a key reason that they are using Linode and plan to grow their use of Linode is because of these deployments. And the advantage of the lighter-weight deployments is we can get into markets, into regions where it's hard to build out a massive core compute data center so that we're going to have before -- next year more than double by having over a couple of dozen of the core compute data centers, but several dozen more of these lighter-weight distributed locations, where you can do the compute.

You wouldn't have huge monolithic storage there, but you don't need that. That can be in the core regions. And that is a key consideration to some of our larger customers that are working with Linode today, doing proofs-of-concept, so in some cases, already running mission-critical applications. Because those lighter weight regions being closer to enterprise data centers in many parts of the world and to end users gives you better performance.

And I think that will put Akamai in a great position to have equal or better performance than the hyperscalers. Of course, we have already our edge deployment with 4,000 locations, which also supports edge computing on top of delivery. And for a total -- lower total cost of ownership. So yes, the lightweight distributed regions are important for a lot of our customers and prospects with among the major enterprises on Linode.

[Operator]

And our next question today comes from Michael Elias at Cowen and Company.

[Michael Elias, Cowen]

The first one, you mentioned it a bit ago relating to your long-term guidance and M&A on the security front. Just a question around how would you describe the pipeline of opportunities for M&A in the security market? And as part of that, maybe any capabilities which are top of mind as you think about adding to the platform?

[Thomson Leighton, Co-Founder, CEO & Director]

Yes, we have a large pipeline, and we generally do. We're constantly looking for appropriate acquisitions. As Ed said, we're very disciplined buyers, though. And the market as a whole is still highly priced. I think the realities of what's going on in the global economy haven't fully set in yet. That may take another year. And so because we're very careful buyers, we're being very selective there.

I think there's a variety of capabilities that would be interesting as tech tuck-ins. And occasionally, we'll make an acquisition with a product adjacency. I think Guardicore has been a fabulous acquisition. They're the market leaders now in segmentation, making Akamai the market leader there. And I think that's the most important defense an enterprise can have.

You can buy every company's Zero Trust offer, and malware is still getting into enterprises. And the real key is to identify it quickly and proactively block it from spreading. And that's how you limit the damage caused by ransomware and data exfiltration attacks. And that's what Guardicore does. And so I think very important strategic product with enterprise security and Zero Trust. But over time, I think there'll be other capabilities that we'll be interested in, in terms of broadening the portfolio.

[Michael Elias, Cowen]

Got it. Now just a philosophical question for you, Tom. Over the years, you've taken steps to continue to grow the business and expand Akamai into new verticals. But the stock really hasn't responded in the way that I think you would have liked, just given some of your prior comments. My question for you is, as you think about executing the long-term vision for Akamai, do you believe being a public company is the right setting for you to achieve that long-term vision?

[Thomson Leighton, Co-Founder, CEO & Director]

Yes, sure. I think it's great being a public company. And yes, I do think the stock is undervalued where we are today in this market. And that's why I'm going to buy more shares. It's -- and over time, the stock has grown. And I think there's excellent prospects for future growth as we continue to grow Akamai.

And I'm really excited about what we can do in the compute landscape. That's an enormous market. You just look at Akamai, next year, probably security will be our biggest product line. That's a big step forward, given that we started as a CDN company. And I think if you look 3 to 5 years down the road, well, maybe in that time frame, compute could be our largest product line.

So I think there's lots of opportunity for continued growth. We're very disciplined when it comes to cost, and that means there's a lot of opportunity for bottom line growth in earnings per share. We've continued to buy back our equity to reduce the number of shares outstanding. So I think there's a great value proposition for public Akamai shareholders.

[Operator]

And our next question today comes from Tim Horan with Oppenheimer.

[Tim Horan, Oppenheimer]

I hate to harp on it, but the sales slowdown, can you just give us a little more color maybe when you started to see it? Is it continuing? Maybe what the lag is in terms of sales and revenue showing up? Just I'm trying to get a sense of what next year's revenue growth could be. I guess, at a high level, are we looking at 2 or 3 more quarters of flat from what we know now? Or at a high level, can growth be better next year than this year at this point?

[Edward McGowan, CFO]

Jim, let me take a stab at that. So I would say we started to see it really in this quarter and continue -- sorry, this quarter, meaning Q3, sorry, we're reporting Q3, continuing here in Q4. Hard to say how long this economic slowdown lasts. Now keep in mind, most of our business is under contract.

The impact of the new signings doesn't have an overly material impact on the business, especially in any one given quarter. Obviously, over a prolonged period of time, it can slow growth down a bit. I think the bigger thing to think about as you build your models is the impact that foreign exchange has had. I've been trying to call it out as we go. Obviously, the dollar has gotten stronger throughout the year.

So when you think about from an as-reported perspective, that's going to be a pretty big headwind. If you annualized at a point we went back all the way to the beginning of the year, you're talking a couple of hundred million dollars of revenue or $0.40 of EPS, a couple of points of operating margin.

But that's a much bigger issue in terms of growth. And given our strong growth internationally, FX, assuming the dollar continues to get stronger, is something that I'd be more concerned about from a growth perspective. But we'll give you an update on guidance next year when we have our Q1 call -- or excuse me, our Q4 earnings call in Q1. So I'm not going to provide any guidance right now, but hopefully that gives you enough color to think about it.

[Tim Horan, Oppenheimer]

On FX, some of your largest competitors, particularly in cloud, but even on the kind of CDN security space, people had a bundling kind of charge in dollars pretty regularly. And some of your other smaller, one in main competitors in Linode has raised prices quite a bit here lately. Have you thought about maybe switching over the pricing in dollars? Or do you do much of that? And have you taken any pricing steps to increase prices?

[Edward McGowan, CFO]

Yes. So we tend to price most of our -- not all of our international business is in the local currency, most of it is, though. So you can see that's why we have such a big impact. Generally speaking, it's hard to switch with a customer who has been paying in one currency and then switching to another one.

Then you also -- the -- in terms of price increases, too, that's not something that we're considering at this point. Obviously, it's kind of a risky thing to do, when you see companies raise prices. Part of the reason we're actively looking to move our cloud spend is for the reason that we're seeing the suppliers in that area either increase price or not give any commensurate declines with volume increases.

So I think the long-term strategy of introducing price increases can come back and backfire on you. So we're not planning on -- we're not going to be making any changes in terms of changing customers out from paying in local currency to dollars.

[Tim Horan, Oppenheimer]

In the U.K., you're 20% below your peers in the last 9 months of price reduction effectively. But my last question is, Linode, is your OpEx and CapEx run rate enough to transition Linode and grow Linode?

[Edward McGowan, CFO]

Yes. So if you think about where the investments are going to be, and that's where we're primarily investing our headcount is in Linode and also in security. And Tom also mentioned that we'll be moving some of the people that have skills that are transferable. If you think about building out, scaling up a CDN, there's a lot of transferable skills. So we'll be able to move some folks that have talent into that group as well.

So that won't put any pressure on the bottom line, but we will be spending some money and continuing to grow because we think the opportunity is significant. And then on a CapEx perspective, we will be building out to take in the consideration, moving our own workloads as well as the future demand, and we'll give you an update on that at the next call.

[Operator]

And the next question comes from Amit Daryanani with Evercore.

[Amit Daryanani, Evercore ISI]

I have two as well. I guess maybe on the first one, if you could just talk about the edge business, the Linode asset. And I guess, maybe the question I struggled with a fair bit is, can you grow this business on the cloud infrastructure side without sacrificing operating margins over the next several years? Or is that growth in Linode on the edge side going to come at lower margins inherently?

[Edward McGowan, CFO]

Yes. So good question. I think we bring some pretty interesting synergy. I just mentioned on the last question how we've got a lot of skill sets in-house that can do, say, network build-outs, for example. We don't have to build out a separate team to do network build-outs. We've got engineering talent in-house.

We also have a big enterprise sales force in place that Tom talked earlier in one of the earlier questions about the spending of some of our larger verticals and the relationships we have with those customers who are spending probably 10 to 15x more on cloud computing than they are on CDN. So there's a significant synergy that you get there.

And then also with our network infrastructure that we have built out, Tom talked about connecting our backbone to the existing Linode centers. That drives a significant cost benefit. So I think that actually, we could have very attractive operating margins, similar to what I showed on the IR Day. We get pretty good operating leverage, just like we did with the security business. So long term, our goal is to get back to 30% or higher in operating margin. And I think as we scale that business, we should be able to do it.

[Amit Daryanani, Evercore ISI]

Got it. And then, I guess, maybe I'd just stick to that theme around Linode. Are there certain use cases that make Linode more attractive more so the top 3 cloud providers that are out there? I guess I'd just love to understand, when you folks walk into a customer pitch, what are the reasons want to use Linode versus some of the peers that might provide a cloud solution at least cheaper? Maybe that would be really helpful. Like what are the 2, 3 reasons that you think stands out?

[Thomson Leighton, Co-Founder, CEO & Director]

Yes. Let me answer that question in the context of where we'll be next year. Because as you know, we're doing a lot of the build-out now. We're doing -- adding a lot of the functionality now. But if you look at where we'll be this time next year, I think, first, our footprint will be better than that of the hyperscalers. We'll be closer to a lot more enterprise data centers' end users.

So that means better performance. I think we'll be hooked in now then to the Akamai platform with 4,000 POPs to do delivery, to do the outer layer of security, to do edge computing. So that's a big advantage. As Ed noted, that also lowers our costs substantially for egress. And so lower total cost of ownership is another, I think, very attractive feature that Akamai would have.

Akamai is known for scalability, for reliability in addition. And there's been some pretty well-publicized issues with some of the hyperscalers in terms of reliability, extended issues. And I think that's an area where we can be very competitive. Really the only area where we wouldn't be as competitive as in the large number of third-party apps in the ecosystem that are available as managed services on the hyperscalers. And that's a key way that you get vendor lock-in.

And so customers that want -- enterprises that want to have that, fine, okay? But if you don't want lock-in and you do want better performance at a lower cost, I think Akamai will be very competitive. And also, it's good to keep in mind just relative scale. Those hyperscalers are giant companies. If we can go get 1%, 2% market share in a multi-hundred billion dollar business, that will be very meaningful for us. And I think we're in a position that we can go do that.

One other issue is that in the sectors where we're very strong like media and commerce, those companies, they compete heavily with at least some of the hyperscalers. And with the cost of the hyperscalers rising, that's becoming more of an issue for those companies.

You're paying a large bill to a company that's buying out the media rights from underneath you. And that's sort of tough for some of them to take. So I think that also gives us a competitive advantage. We don't compete with our customers. We will help them grow with their business and be good partners to them and they can trust us.

[Operator]

And our next question today comes from Mark Murphy at JPMorgan.

[Sonak Kolar, JPMorgan Chase & Co.]

Sonak Kolar here on for Mark Murphy. Tom, can you get a bit deeper on the new pricing strategy, particularly around delivery? Have you noted any incremental changes in customer retention or churn levels as a result of some of these pricing adjustments, specifically around some of the inflationary environment pressures driving price sensitivity in the market?

[Edward McGowan, CFO]

Yes. This is Ed. No, we haven't seen anything notable. Like if anything, we're actually starting, like I said, to see some bit of a moderation in the pricing declines. So we haven't seen anything on the churn side.

[Sonak Kolar, JPMorgan Chase & Co.]

Got it. That's very helpful. And then a quick follow-up. Would you say some of the macro pressures that you've called out have intensified in Q3 relative to Q2? Or has it remained fairly in line with some of the pressures that you called out during the last earnings call?

[Edward McGowan, CFO]

Yes, I'd say it's intensified a bit in Q3.

[Operator]

And our next question today comes from Rudy Kessinger with D.A. Davidson.

[Rudy Kessinger, D.A. Davidson & Co.]

I don't want to belabor the point on security growth. But again, when we exclude Guardicore and look at it at constant currency, it looks like organic has come down about 5 points from Q1 to Q3. And so if you were to look at it, could you maybe break out like what's been the impact in your assessment from the macro and -- versus just maybe some of the larger products maturing and growing slower that's led to that roughly 5-point deceleration in the last couple of quarters?

[Edward McGowan, CFO]

Yes, I'd say it's a tough one to really figure out what the impact is on the macro. I'd say that's probably -- maybe that's a point or 2. But I think it's really the issue of what Tom talked about, where if you look at the biggest product, we're the market leader in web app firewalls growing faster than the market, but that market isn't growing as fast as some of the other markets that we're in, and it's just not at scale yet.

So I'd say that's the bigger issue, except those newer products are growing faster at the scale that we're at, just isn't offsetting the slower growth in those -- in the bigger products.

[Rudy Kessinger, D.A. Davidson & Co.]

Okay. And then on Linode, I don't know if you gave it. Could you share how much revenue Linode did in the quarter? And then last quarter, given the commentary today about the increase in cost of the hyperscalers, last quarter, you talked about moving your hyperscaler spend over to Linode internally. Have you started that process yet? And if not, when do you plan to do so?

[Edward McGowan, CFO]

Yes, I'll take the first one.

[Thomson Leighton, Co-Founder, CEO & Director]

Go ahead, Ed. You take the first part and I'll...

[Edward McGowan, CFO]

I'll do -- mine is pretty simple. So Linode added about $33 million this quarter. Tom, why don't you talk about the movement?

[Thomson Leighton, Co-Founder, CEO & Director]

Yes. So that's well -- the migration of our cloud spend to Linode is well underway. We've already done some. The lion's share of that work or the migration will take place, I would say, over Q1 to Q3 next year. And by the end of next year, we ought to have the vast, vast majority migrated.

[Operator]

And ladies and gentlemen, our next question today comes from Tom Blakey with Truist Securities.

[Tom Blakey, Truist Securities]

I think it's been touched on a couple of times by my peers here. Just wanted to get -- go back to that. Some sort of normalized CapEx level, from my estimation, we're clearly overspending on Linode as a percentage of Linode revenue, anyway, right, if you do the percentage of total revenue. And you've made great strides in terms of lowering the CapEx from a delivery perspective down to this kind of 3% to 4% range.

But as things kind of normalize, Tom, when you look out, and we're at , maybe 2/3 of revenue coming from compute security, CapEx is nil. You're just riding the rails of what already exists. What does the CapEx kind of bridge or normalized structure of this company look like a few years from now?

[Edward McGowan, CFO]

Yes. So I'll take a stab at that. Tom, if you want to add anything, feel free to jump in. So obviously, with the -- what we're talking about here, where Linode was primarily focused on small, medium business, and we're moving towards the enterprise workloads, you're talking about much larger workloads. Even with our own spend and what we're moving, Tom talked about on the last call, we had that we're spending roughly $100 million, that's growing pretty fast.

That's a much bigger scale than what Linode was doing. So looking at CapEx as a percentage of existing Linode business isn't the right metric to look at, at this point. But think of it as we're in the build phase. So you can see CapEx is starting to creep up a bit as we're building out, as we see better visibility into demand and also as we start to build out our plans to migrate the workloads that we do have on the hyperscalers on to us.

So you're going to see a bit of a disjoint. So if you're looking at that as a metric, it doesn't send the right signal. I'd say look at that as more of a bullish signal in terms of how we feel about our customer -- pending customer demand and also how we think we'll be very successful in being able to migrate those workloads. So for the next, say, several quarters, you're going to see more CapEx going into Linode and then the revenue and the cost savings will follow.

But then if we get -- it really depends on the growth, right? So as you're growing revenue at significant rates, and so you're doubling revenue, you're going to obviously have a higher percentage of CapEx. But then at some point, it will normalize out to approximate what the future revenue growth would be. And so, let's say, for example, we get to 20% growth as a run rate in the long term, your CapEx will probably be somewhere in that range.

[Tom Blakey, Truist Securities]

I'm sorry, I didn't understand the last comment of 20% growth in Linode, the CapEx was...

[Edward McGowan, CFO]

No, no, no. So I was using that as an example. What I was saying is we'll give you detailed guidance on what we're going to do next year, but we're in a building phase now where Tom talked about getting into many new centers we're building out for our own demand and also what we're hearing from our customers. So what I was saying is if you're looking at CapEx as a percentage of Linode, it's not the right way to be thinking about it because we're going after a much different business, right? We're going after big enterprise workloads.

So there's a build phase where you have to build out ahead of the demand. And then you'll start to see the revenue come our way. And as you get to scale, like many years out when you get to scale, you can start thinking about as a proxy that roughly speaking, your CapEx would approximate what your future demand is. So if you, say, have a long-term run rate of 20% or 30%, your CapEx will be somewhere in that range. But in the near term, it'll be higher than...

[Tom Blakey, Truist Securities]

That's exactly what I was asking. I understand, obviously, you're overspending today. And that comment was just a structural comment for the question on what CapEx would be as a total percentage of revenue, total revenue in the -- again, when you're at scale for Linode? Will it be structurally lower or higher than delivery?

[Edward McGowan, CFO]

We'll know when we get there. There'll probably be -- certainly higher than what we're seeing in delivery now, but it is a more capital-intensive business by nature.

[Tom Blakey, Truist Securities]

It's a more capital-intensive business with compute, okay.

[Operator]

And our next question today comes from Will Power at Baird.

[Charles Erlikh, Robert W. Baird & Co.]

This is Charlie Erlikh on for Will. I just wanted to ask a 2-parter on the comment that you guys are going to basically take some resources from delivery and put them into security and compute headcount-wise and hiring-wise. So the first part is, how do you feel about competition for talent in the security business and the compute business maybe relative to a few months ago? Kind of what does that hiring environment look like in those 2 businesses?

And then part 2 on the delivery side, how should we interpret that comment as far as trying to turn around that delivery business and just sort of what should we expect from that business as far as trends going forward?

[Thomson Leighton, Co-Founder, CEO & Director]

I'll take the first question. It's still a competitive market for hiring. We've been very pleased to see our attrition rates take a major drop over the last quarter. We had stayed at low attrition rates through COVID, well better than market, ticked up a little bit in the first half of the year, but now again down to, over the last 3 months, very low attrition.

We have very successful recruiting. Akamai is considered to be a great place to work as measured by the various studies that are done and also employee satisfaction surveys that we do. So people really like working at Akamai. We have really great employees, very smart. We set a high bar for who we recruit.

And of course, everybody wants to hire those people. And pretty much everybody wants to hire Akamai employees. But our retention rates are good. I would say our success in hiring is good, but it's a competitive market out there. And Ed, do you want to talk about the delivery business?

[Edward McGowan, CFO]

Yes, sure. So the way to think about the delivery business, I'll call on 4 factors in terms of thinking about, as you called it, a turnaround. Obviously, one is renewal. So we went through a very heavy phase of renewals. So we're not going to have that next year. We'll always have some renewals, but it's very unusual to see 8 of your top 10 customers renewing at the same time.

Number two is pricing. So we're -- as we talked about several times on the call, we are moderating the discounts that we provide on pricing. And then the third thing is traffic. Traffic, we're starting to see some encouraging signs, albeit early, that the Internet has been growing at 30% a year for many, many years. This year is a sub-30% year, but it's reasonable to think that we should start to get back to more traditional growth rates.

And then the fourth thing I would say is I think we get a tailwind in our delivery business from being in the compute space. We actually do see some customers that are on hyperscalers get to a certain size that they -- even though they offer their own CDNs, come to us for better performance. So as we add customers, get into new verticals, et cetera, we do have the opportunity to just grow the delivery business by being a proxy of being in the compute business.

[Operator]

And our final question today comes from Alex Henderson at Needham & Company.

[Alexander Henderson, Needham & Company]

Sliding in before the final. Nice. So I wanted to go back to the commentary that you've made about the outlook for the upcoming quarter, particularly in the security space. If I adjust the numbers for the contribution from the acquisition of Guardicore, I'm getting an as-reported growth rate of around 9%. And I'm wondering, given your commentary about more difficult conditions and a little larger currency translation year-over-year in the fourth quarter, whether in fact you're expecting the security business to slow to that level in your guidance?

[Edward McGowan, CFO]

Yes. Alex, this is Ed here. I'll give that one a try. Obviously, we don't break out specific guidance for individual products like that in the quarter. But right now, the FX impact, as you quoted an as-reported number, is about 6%. So as-reported was 13%, constant currency was 19%. So if you assume that, you're getting to about -- if you're using 9%, you get to about 15%.

Since we're lapping, in constant currency that is -- since you're lapping the Guardicore acquisition at this point, and we just came off a 15% organic growth rate, if you back up the contribution from Guardicore in Q3, that's probably a reasonable number.

If you're solving for what we gave you in terms of 20% constant currency, somewhere in that 14% to 16% constant currency. Obviously, I can't predict where FX is going to be. But if you just kind of keep it what it was this quarter, that's -- you're doing roughly the right math.

[Alexander Henderson, Needham & Company]

So a similar kind of question. If I take the Linode comments, it looks like that actually accelerated from about a 15% baseline growth rate to around 20%, making the adjustment for the Linode acquisition. On an as-reported basis, that's actually pretty good, considering the currency. Can you talk a little bit about the geographic play between -- in the Linode business and how much currency would have impacted that side of it?

And then again, as we look into the baseline growth rate, it does seem like it's accelerating. Can you talk a little bit about whether that's a function of the investments you're making in marketing and like? Or whether that's something that's sustainable in the guide?

[Edward McGowan, CFO]

Sure. So I'll start with the currency with Linode. So not -- they -- when we acquired Linode, most if not all of their revenue was in U.S. dollars. So they were not billing in local currency. So there's not as much currency headwinds associated with the Linode portion of the business. Obviously, with the non-Linode compute business, we have the normal dynamics in our business.

On the investments in terms of marketing, et cetera, I think we can maintain the current levels, and I'm not expecting a significant increase in marketing spend next year. We are increasing it a bit, but nothing significant. From a sales perspective, we are adding some sales folks, some specialists, especially on the technical side, to help our sales force, but it's not going to be a significant investment there.

We believe our sales force can be trained, and we're also changing our comp plans to have an added incentive for compute. So all of that, I think, can fit into sort of a normal run rate expense level that we've been sort of operating, and I don't see any major investments in that part of the business.

[Alexander Henderson, Needham & Company]

If I could slide in one last one, since I'm the last guy here. As the mix shifts here, how do you expect the mix shift between the segments to start impacting the overall margins?

[Edward McGowan, CFO]

Yes. So obviously, if you go back to the IR Day slides, and there was a question earlier about the leverage that we get on the compute business, as the faster-growing parts of the business security and compute become a bigger part of the business, we should get some operating leverage. It won't happen right away, but over time. Then obviously, there's the dynamic that we talked about with some accelerated CapEx ahead of revenue. So that's another thing to keep in mind.

The other thing you notice -- I talked a bit about having a bit of a pressure on the gross margin line. That also, I would say, is more of a temporary thing. We will be reducing the gross margin line as we move our third-party costs over, but also with some of the build-out costs, including some of our colocation agreements, network build costs, a little bit that's front-loaded.

So as revenue scales, we should start to see some scale in the gross margin line as well. So as Tom talked about, security should be our biggest product line. It's got higher gross margins, higher operating margin. So we should start to see that flow through as the mix changes to those 2 product lines over time.

[Thomson Leighton, Co-Founder, CEO & Director]

And thank you, everyone. In closing, we will be presenting at a number of investor conferences and presenting at events, road shows and other things throughout the rest of the fourth quarter. Details of these can be found in the Investor Relations section of akamai.com. So thank you for joining us. And all of us here at Akamai wish continued good health to you and yours, and have a nice evening.

[Operator]

Thank you, sir. This concludes today's conference call. We thank you all for attending today's presentation. You may now disconnect your lines, and have a wonderful day.