Akamai Technologies, Inc. (AKAM) Q4 2022 Earnings Call Transcript

Akamai Technologies, Inc. (NASDAQ:AKAM) Q4 2022 Earnings Conference Call February 14, 2023 4:30 PM ET

Company Participants

Tom Barth – Head, IR

Tom Leighton – Co-Founder, CEO & Director

Ed McGowan – CFO

Conference Call Participants

James Fish – Piper Sandler

James Breen – William Blair

Keith Weiss – Morgan Stanley

Tim Horan – Oppenheimer

Mark Murphy – JPMorgan

Rishi Jaluria – RBC

Amit Daryanani – EVercore ISI

Fatima Boolani – Citigroup

Frank Louthan – Raymond James

Rudy Kessinger – D.A. Davidson

Ray McDonough – Guggenheim Securities

Michael Elias – Cowen & Co.

Operator

Good afternoon, and welcome to the Akamai Technologies, Inc. Earnings Q4 Fiscal Year 2022 Earnings Conference Call. [Operator Instructions] Please note that this conference is being recorded.

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

Tom Barth

Thank you, operator. Good afternoon, everyone, and thank you for joining Akamai’s fourth 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 Form 10-Q. The forward-looking statements included in this call represent the company’s view on February 14, 2023. Akamai disclaims any obligation to update these statements to reflect new information, future events or circumstances, except as required by law.

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

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

Tom Leighton

Thanks, Tom, and thank you all for joining us today. I’m pleased to report that Akamai delivered strong results in the fourth quarter, exceeding the high end of our guidance range on both the top and bottom lines, despite ongoing challenges with the global economic environment. Q4 revenue was $928 million, up 6% year-over-year in constant currency. Our revenue growth was driven by continued strong demand for our security products, our fast-growing compute business and by higher-than-expected delivery traffic. Security and compute accounted for 55% of our overall revenue in the fourth quarter and grew a combined 22% year-over-year in constant currency.

Non-GAAP operating margin in Q4 was 28%. Q4 non-GAAP EPS was $1.37 per diluted share, down 2% year-over-year in constant currency. For the full year, revenue was $3.62 billion up 8% over 2021 in constant currency. Non-GAAP operating margin was 29%, down from 32% in 2021 and slightly below our goal of 30%. The decline last year was due to the impact of foreign exchange, the challenging macroeconomic environment, the investments we made to grow Guardicore’s segmentation product and the rising cost of third-party cloud services.

As Ed will describe in a few minutes, we’re taking several actions to reduce costs and to shift resources to areas with the strongest potential for growth, such as cybersecurity and especially cloud computing. Going forward, we anticipate that our margins will likely remain slightly under 30% in the near term, and our goal is to grow margins back over 30% during the medium to long term. Non-GAAP EPS last year was $5.37, down 1% over 2021 in constant currency. 2022 was another strong year for cash generation at Akamai, with $816 million in free cash flow, representing 23% of revenue.

Akamai’s strong cash generation enables us to make strategic acquisitions while also returning value to shareholders. In 2022, we spent $608 million to buy back 6.4 million shares. Over the last 10 years, we’ve reduced the number of Akamai shares outstanding by approximately $21 million or 12%.

I’ll now say a few words about each of our three main lines of business, starting with security. Our security products generated revenue of $400 million in Q4, up 14% year-over-year in constant currency. For the full year, security revenue reached $1.54 billion and grew 20% over 2021 in constant currency. We saw especially strong growth through our market-leading Guardicore Segmentation product with revenue reaching $68 million for the full year.

New segmentation customers in Q4 included one of the largest insurance companies in the U.S., a leading Internet services conglomerate in Japan and one of the largest banks in Scandinavia. Enterprises are choosing our segmentation solution because of its ability to protect against ransomware and data exfiltration attacks and also for the visibility it provides into their internal infrastructure. These are also among the reasons that Akamai was named as the Leader in The Forrester New Wave for Microsegmentation last year. We also saw large wins for our market-leading application security solutions in Q4, including at one of the UK’s largest multinational energy companies, one of the big three multinational banks in Singapore and two of the largest tech hardware companies in the U.S.

Overall, security accounted for 43% of our revenue last year, up from 39% in 2021. In 2023, we expect security to become our largest line of business. This represents a significant milestone in our evolution since we pioneered the CDN marketplace 25 years ago. That said, and as the security business becomes larger and with customers becoming more cost conscious due to the challenging macroeconomic environment, the growth rate of our security business has slowed, a trend that we anticipate will continue over the coming year.

As you might expect, we’re working hard to realize the full potential of our security business, both in terms of growth and efficiency. For example, we’re in the process of moving the compute components of our security products from third-party cloud providers to our new Akamai connected cloud platform a transformation that will save us a substantial amount of OpEx over the next several years. We’re redeploying resources within security from lower growth areas to high-growth areas such as segmentation. And we’re redeploying go-to-market resources to achieve stronger cross-sell and penetration within our existing base.

We’re also working more closely with partners to drive better adoption among new customers. And we’re continuing to innovate new capabilities, such as our recently released Account Protector and our new Brand Protector solutions to keep our customers safe amidst a rapidly evolving attack landscape.

While we believe that our security business will continue to generate strong returns for our shareholders, we foresee an even bigger opportunity in cloud computing and its potential to return Akamai to double-digit top line growth over the longer term. Our compute business performed well in Q4 with revenue of $112 million up 65% year-over-year in constant currency. For the full year, compute revenue was $405 million and grew 64% over 2021 in constant currency.

Earlier today, we unveiled Akamai Connected Cloud, our massively distributed platform for cloud computing, security and content delivery. Akamai Connected Cloud links Linode’s 11 core data centers with Akamai’s 4,100 edge computing locations. In addition, we’re in the process of building out 14 more core enterprise scale data centers with at least 3 expected to come online in the next few months. We believe integrating these core cloud computing data centers with our unique edge platform will allow us to offer customers better performance, greater scale and lower cost for enterprise workloads.

We also plan to have our new virtual private cloud capability and the first of more than 50 distributed cloud computing sites available in the second half of the year. The distributed sites will enable us to bring cloud computing much closer to end users around the world, which will further enhance the performance benefits of Akamai Connected Cloud.

Of course, and as Ed will describe shortly, we’ll be incurring substantial CapEx and colocation costs associated with the build-out of our compute infrastructure over the near term. We’re also in the process of recasting approximately 1,000 positions or about 10% of our workforce to spend the majority of their time working on the development, deployment, support and go-to-market efforts associated with Akamai Connected Cloud. Because of the natural synergy and close integration between cloud computing and our existing edge platform, we believe we can accomplish this transformation without adding significant headcount to the business.

This shift will also further enhance the efficiency of our delivery business. We’re undertaking this ambitious investment in Akamai Connected Cloud because we believe it will create substantial value for shareholders in the medium and long term. We expect to achieve nearly $0.5 billion in revenue from compute in 2023, and the investment we’re making this year should help drive that number substantially higher in 2024 and beyond as we use the new capabilities and capacity to support mission-critical enterprise workloads.

I think it’s worth noting that Akamai is taking a fundamentally different approach to the cloud computing market than providers who base their platforms solely on core data centers. Our strategy is to offer the world’s most distributed platform placing compute, storage, data base and other cloud services closer to end users and enterprise data centers. As IDC’s VP of Research Dave McCarthy says “The cloud’s next phase requires a shift in how developers and enterprises think about getting applications and data closer to their customers. It redefines how the industry looks at things like performance, scale, cost and security as workloads are no longer built for one place but are delivered across a wide spectrum of compute and geography.” IDC adds that “Akamai’s innovative rethinking of how this gets done and how it is architecting the Akamai Connected Cloud, puts it in a unique position to usher in an exciting new era for technology and to help enterprises build, deploy, and secure distributed applications.”

We couldn’t agree more, distributed applications require a distributed architecture. Akamai’s leadership positioned at the edge of the Internet enables us to scale just about everything we touch. We scale content putting digital experiences closer to users than anyone. We scale cybersecurity, keeping threats father away from business and people. And now, we are building on Akamai’s 25 years of experience with scaling and securing the Internet for the world’s largest enterprises, so we can scale cloud computing and provide better performance at lower cost.

Although we still have much work to do, we are encouraged by the reaction from customers who want to realize the value of our approach. Last quarter, a well-known digital fitness platform brought business to us that they previously did with a major cloud provider. They chose Akamai Connected Cloud because we can optimize their performance and provide better economics. When a gaming company suffered a DDoS attack that took out their Internet relay chat servers, they turned to Akamai Connected Cloud to get back on line. After utilizing connected cloud for a few weeks, they also migrated their peer-to-peer match making servers to Akamai. This is what they say to other gaming businesses with similar use cases: “Our adoption of Akamai’s cloud computing services was painless and turnkey. Akamai has a great backbone network and the connected layer between our global servers has been rock solid. With Akamai’s extensive global network we provide a better experience to our gamers by delivering from the edge and reducing latency. With Akamai, there is no reason to go anywhere else.”

As you can see from this example, there’s a strong synergy between our emerging cloud computing business and our delivery and security businesses, especially for customers in the gaming, media and commerce verticals.

Turning now to content delivery. Our CDN business generated revenue of $415 million in Q4, down 8% from Q4 and 2021 in constant currency. For the full year, delivery revenue was $1.67 billion, also down 8% year-over-year. Traffic on the network was better than expected in Q4, reaching a new peak record of 261 terabits per second on December 14 as we supported more than 50 customers globally in delivering the World Cup, along with other streaming, gaming and software download businesses. This World Cup was the first time in Akamai’s 25-year history when we delivered more than an exabyte of data for an event. How much is an exabyte? It’s 1,000 petabytes. That’s 1 billion gigabytes. For the person transcribing this call, that’s 1 byte with 18 zeros after it. That’s a lot of zeros and a lot of traffic. I doubt if anyone has managed such a feat before.

Once again, Akamai finished the year as the CDN market leader by far, as we continue to support the world’s leading brands by delivering reliable, secure, high-performing online experiences.

Looking back at 2022, we’re pleased that we continue to grow the business and add significant new capabilities in the face of serious global macroeconomic challenges. Today, we’re redefining our future with Akamai Connected Cloud to become the world’s most distributed cloud platform with leading solutions for delivery, security and cloud computing. With our expanded strategy and business model, we believe that we’re on a path to provide even greater value for shareholders and to make Akamai the cloud company that powers and protects life online.

Now I’ll turn the call over to Ed for more on our Q4 and full year results and our outlook for 2023. Ed?

Ed McGowan

Thank you, Tom. Today, I plan to provide brief highlights of our strong Q4 results, some color on 2023 and touch on some items to help you with your models and then close with our Q1 and full year 2023 guidance.

Starting with Q4 highlights. We were very pleased with our strong Q4 results despite continued difficult macroeconomic conditions. Q4 revenue was $928 million, up 2% year-over-year or 6% in constant currency. We saw very strong growth in both our compute and security businesses as well as better-than-expected traffic in our delivery business during the fourth quarter.

As Tom mentioned, our compute business was $112 million, growing 61% year-over-year as reported and 65% in constant currency. We continue to be very pleased with the initial feedback from our customers on our future compute capabilities, and we are very optimistic about capturing a meaningful share of our customers’ cloud spend in the years to come.

Our security revenue was $400 million, up 10% year-over-year and up 14% in constant currency. Our delivery revenue was $415 million, which declined 12% year-over-year and 8% in constant currency. Traffic exceeded our expectations during the quarter, led by higher video traffic, stronger-than-expected commerce traffic and record-setting World Cup online viewership.

International revenue was $445 million, up 4% year-over-year or up 12% in constant currency, representing 48% of our total revenue in Q4. Foreign exchange fluctuations had a negative impact on revenue of $2 million on a sequential basis and negative $36 million on a year-over-year basis. Non-GAAP net income was $216 million or $1.37 of earnings per diluted share, down 8% year-over-year and down 2% in constant currency, but $0.07 above the high end of our guidance range.

Moving to our capital allocation strategy. During the fourth quarter, we spent approximately $178 million to buy back approximately 2.1 million shares. For the full year, we spent approximately $608 million to buy back approximately 6.4 million shares. We ended 2022 with approximately $1.2 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.

It’s worth noting that in addition to offsetting dilution, we have reduced shares outstanding by approximately 21.2 million shares or 12% since January 1, 2013.

Before I move on to guidance, there are several items that I want to highlight to help you with your 2023 models. The first relates to a change in our network server useful lives. As some of you may recall, we announced on our Q4 2018 earnings call that we were required to extend the useful life of our network servers from 4 years to 5 years based on the actual server useful life trends. We carefully monitor the useful lives of all of our capital assets annually. And based on the outcome of our most recent review, we now are extending the useful lives of our servers from 5 years to 6 years.

Similar to when we made the change 4 years ago, this extended useful life is a direct result of the continued software and hardware initiatives that we have put in place to manage our global network more efficiently. Because we are now using the servers in our network for an average of 6 years we are required under GAAP accounting to adjust our useful life policy to 6 years beginning in Q1 of 2023. Please keep in mind that this change has no impact on cash flow, but will result in a depreciation benefit of roughly $56 million in 2023 and approximately $31 million in 2024. We have provided a supplemental table in the Investor Relations section of our website that details the impact of this change.

Second, we are expecting non-GAAP gross margins to decline by approximately 2 points in 2023 due to two primary items. First, as we build out our new compute locations, we are required to account for our colocation leases under GAAP accounting standard ASC 842. In order to achieve more favourable unit economics, we often sign longer term colocation agreements that include certain financial commitments. ASC 842 requires we straight line the cost of those future financial commitments over the life of the agreement. As a result, we expect to record approximately $40 million of noncash colocation costs related to this accounting standard in 2023. The second item impacting gross margin is our third-party cloud costs, as we’ve mentioned in the past, we expect to migrate the majority of our third-party cloud spend on to our own cloud infrastructure over the next 12 to 18 months. That said, we expect the majority of the migration effort to impact the back half of the year. We expect we will incur just over a $100 million of third-party cloud costs and cost of goods sold in 2023.

We do, however, expect we will exit the year on a path to significantly lower our third-party cloud costs from 2024.

Third, we expect international sales will represent nearly half of our total revenue in 2023 and the currency markets remain incredibly volatile. I will provide more detail on the impact of currency on each quarter’s earnings call, but it’s important to note that the strengthening or weakening of the U.S. dollar can have a material impact on our reported results and guidance. As a reminder, the currencies that have the largest impact on our business are the euro, the yen and the Great British pound.

Fourth, I want to remind you of the typical seasonality that we experienced on the top and bottom lines throughout the year. Regarding revenue, the fourth quarter is usually our strongest quarter, and we typically see a step down in Q1 revenue from Q4 levels. Regarding profitability, as a reminder, in Q3, we have the annual company-wide merit increase; and in Q4, we typically see higher sales commission expense.

And one final thought before we move on to guidance. Tom mentioned that we will be investing in what we believe will be two areas of higher growth for the company for years to come: security and compute. While we expect to manage the business below our target operating margin of 30% in 2023 and we expect 2023 to be a higher-than-normal year for CapEx, we will continue to reduce costs and drive efficiency gains in key areas such as third-party cloud savings. We expect to save over $100 million in annual cost as we migrate workloads from the hyperscalers to our own platform over the next 12 to 18 months.

Real estate rationalization. As our employees have largely elected to work remotely, we expect to reduce our real estate costs by approximately $20 million in 2023 and achieve further savings in 2024.

Shifting resources. As Tom mentioned, we’ll be incurring substantial CapEx and colocation costs associated with the build-out of our compute infrastructure over the near term. As a result, we are prioritizing certain actions and retasking approximately 1,000 positions from other parts of the business to our compute business. In addition, during the fourth quarter, we closed approximately 500 open positions and we will continue to be very prudent with any headcount additions during the year.

Finally, we are lowering CapEx related to delivery. We expect to reduce our CapEx related to the delivery business to 4% of revenue in 2023.

Now moving on to guidance. Our guidance for 2023 assumes no material changes, good or bad, in the current macroeconomic landscape which we view as challenging, but navigable. For the first quarter of 2023, we are projecting revenue in the range of million to $900 million to $915 million or up 0% to 1% as reported or 2% to 3% in constant currency over Q1 2022. Foreign exchange fluctuations are expected to have a positive $13 million impact on Q1 revenue compared to Q4 levels, but a negative $19 million impact year-over-year.

At these revenue levels, we expect cash gross margins of approximately 73%.

Q1 non-GAAP operating expenses are projected to be $299 million to $303 million. We anticipate Q1 EBITDA margins of approximately 39% to 40%. We expect non-GAAP depreciation expense to be between $109 million to $111 million, and we expect non-GAAP operating margin of approximately 27% to 28% for Q1. And with the overall revenue and spend configuration I just outlined, we expect Q1 non-GAAP EPS in the range of $1.30 to $1.34. This EPS guidance assumes taxes of $43 million to $44 million, based on an estimated quarterly non-GAAP tax rate of approximately 17.5%. It also reflects a fully diluted share count of approximately 157 million shares.

Moving on to CapEx. We expect to spend approximately $220 million to $228 million, excluding equity compensation and capitalized interest in the first quarter. This represents approximately 24% to 25% of anticipated total revenue.

Looking ahead to the full year, we expect revenue of $3.7 billion to $3.78 billion, which is up 2% to 4% year-over-year, both in as reported and in constant currency. We expect security revenue growth to be in the low double digits for the full year 2023. We are estimating non-GAAP operating margin of approximately 27% to 28% and full year CapEx is expected to be approximately 21% of total revenue.

We expect our CapEx to be roughly broken down as follows: approximately 4% of revenue for our delivery business; approximately 9% of revenue for our compute business, of which roughly $100 million of that will be for internal workloads moving in-house and the remainder for future revenue growth; approximately 7% of revenue for capitalized software; and the remaining about 1% for IT and facility-related spend.

We expect non-GAAP earnings per diluted share of $5.40 to $5.60. And this non-GAAP earnings guidance is based on a non-GAAP effective tax rate of approximately 17.5% and a fully diluted share count of approximately 157 million shares.

In closing, we are very pleased with the strong finish to 2022, and we are excited about our growth prospects in both security and cloud computing.

Now, Tom and I would be happy to take your questions. Operator?

Question-and-Answer Session

Operator

[Operator Instructions] Our first question will come from James Fish with Piper Sandler.

James Fish

I appreciate all those details on the moving part. Ed, I know there’s a lot there. I actually want to dive into the security business here. Tom, you had mentioned thinking about go-to-market investments and looking to prioritize on how to better address the market. Are you guys changing at all how you’re approaching the market, especially on the network security side with Zero Trust as it seems to be a little bit of a disconnect? And, Ed, is it possible we can actually get an update on what that access control Zero Trust kind of business finished at for 2022?

Tom Leighton

Yes. No, good question. And we are increasing the allocation of our go-to-market investments around the enterprise side and Zero Trust. We have specialist teams there really focused on Guardicore and now the other enterprise products as we integrate them in with the Guardicore solution. And we’ve had a lot of success with that team, as you can see from the really good growth in the Guardicore product. And so we’re doubling down there. And over the course of this year, you’ll see, I think, more cross-sell with our Enterprise Application Access product in Guardicore and ETP. And then I’ll turn the other question over to Ed.

Ed McGowan

Jim. Yes. So what I’d say about Zero Trust is we ended the year on a run rate of about $200 million in Zero Trust. And we had a very strong finish to Guardicore. In Guardicore, we’d expect this year to be on a run rate of $100 million. So a very, very strong finish to the year with Guardicore.

James Fish

Helpful. And just a follow-up on the compute side. A big question we’ve been getting is whether the compute business could get traction really outside the media vertical? Or is that really going to be the focus in terms of gaming and streaming kind of storage and compute on the back end? And kind of what’s the confidence level that you’re not going to see an erosion in that more traditional windows that’s SMB based?

Tom Leighton

I think media is the big first set of adopters. In fact, as you know, one of the world’s largest, if not the largest social media company is already using it. And I think that’s because media really is concerned about performance. And so we’re doing things like transcoding, much closer to the end user. And that’s a key thesis of our architecture and our approach to compute is to be much more distributed with the compute, have containers and VMs much closer to the end users.

Gaming, another example, what that makes a lot of sense. Things like leaderboards, manage groups of users as they play the game, you need low latency, there’s a lot of back and forth with the clients and you want that to be close to the end user.

That said, this is not limited to media. Commerce is also very sensitive to performance, and we’ve already signed up commerce companies. And commerce companies, of course, especially with these macroeconomic challenges. Akamai Connected Cloud, better performance at a more competitive price point. So I think you’ll see penetration in the commerce vertical as well.

And of course, Akamai is really close to the big companies in media and commerce. I have the opportunity to talk to the most senior executives in many of these companies, and they’re very interested in what we can do for them. And they’ve been asking us to do this. And I think they’re excited about the potential Akamai Connected Cloud.

I think down the road, take a little bit longer. The financial vertical is another vertical that’s very strong for Akamai because of our work in security and our market-leading products there. There, we have a little bit more work to do on the certifications. The bar is a little bit higher for financial institutions, but I think that would follow media and commerce in terms of adoption of Akamai Connected Cloud.

Operator

Our next question will come from James Breen with William Blair.

James Breen

Can you just talk a little bit about sort of the internal decision-making around shifting resources to the security and cloud side and maybe a little bit on the delivery side of market share to some extent, giving up some market share within that business sort of better for the overall profitability in the long term?

Tom Leighton

Well, yes, I think it makes good sense to be shifting resources towards the higher growth areas both security but especially in cloud computing. I think we’ll see a lot stronger return on the investment. We still have the market-leading delivery business. We haven’t lost share, to the best of my knowledge. We have turned away some — a small amount of the business that’s very spiky. And that doesn’t make as much sense for us to take that business at the price point isn’t right today. And that’s because traffic growth rates are lower than they used to be. So you have to spend money to build out for the spike. And if your traffic growth rates are lower, it takes longer to fill that capacity on a daily basis.

With delivery, your cost is associated with your peak and your revenue more associated with the daily usage or total aggregate usage. And so we have, as we’ve talked about, turned down some deals there just because the ROI doesn’t make us much sense in this environment, and we get a lot better ROI from investing in compute and I think, obviously, security.

And then within security, we’re reallocating investment decisions there to focus more resources on the fastest-growing security products. So I think it just makes good sense. And for now with compute, it’s something our customers are really asking us to do. And even in this challenging macroeconomic environment, it’s something that I think the timing may be even a little bit better there because of that.

Operator

Our next question will come from Keith Weiss with Morgan Stanley.

Keith Weiss

Excellent. A couple of questions, I have a ton of questions, but I’ll try to limit it to two or three. On the cloud computing side of the equation, the word of the year amongst investors has been optimization, and we’ve heard it from AWS, we’ve heard it from Azure. I’m really interested to hear how that impacted Linode. Given kind of being a lower cost provider in the marketplace, did you guys — did you see optimization of people trying to sort of reduce the amount of consumption on your platform? Are you guys like more of a net beneficiary because of, people are ultimately seeking just kind of lower cost overall?

And then also on cloud, I don’t know, did you guys mention what the contribution was from Linode this quarter? I believe you guys have been given the kind of inorganic contribution over the past couple of quarters. And then I have one for Ed on the margin side of the equation.

Tom Leighton

All right. I’ll take the first part of that and let Ed take the second part. I think Linode is a much smaller company and really sort of a different market than the giant cloud companies that are maybe referring to optimization. So I didn’t see — we don’t see a big impact on Linode. Now what we’re doing is making Linode so it can be used by the big enterprises for mission-critical applications. And that, in part, is building out scale. It’s making it be more distributed, available in more cities, integrating it with the Akamai platform and our Edge regions and increasing the functionality. So a lot of work taking place on the Linode platform so we can sell it at a much higher scale to major enterprises.

And we’re in the stage now with early adopters there. And I think as we get towards the end of this year, we’ll be in a position to take on a lot more business there. And I think the value proposition is that we would have better performance will be more distributed, closer to end users at a more competitive price point. And I think in this environment, yes, pricing matters, especially if you look at — as we talked about, big media, big commerce, they care. And they are spending a lot in the cloud today.

So I think that’s the phenomenon you’ll see take effect as we’re in a position to take on more large-scale business for large enterprises. And then Ed, you can take the second part of that question.

Ed McGowan

Yes, sure. So Linode, it will get more difficult to break up sort of merging in with everything, but it was about $34 million, a little over $34 million. We did see a slight increase in the growth rate as the year went on, and we haven’t seen any change in the consumption around optimization, meaning folks are using less of the cloud at this point.

Keith Weiss

Perfect. Perfect. And then on the expense side of the equation. I just want to make sure I understanding this correctly. It sounds like you’re slowing down or pausing hiring and I’m assuming you close those positions that could be hired into them, but you just like tick it off of the job board, if you will, and then repositioning employees rather than any like restructuring or headcount reduction. So should the way that we should be thinking about it is you’re aiming towards relatively flat headcount in 2023?

Ed McGowan

Yes, Keith. So the way to think it will probably go up a little bit. It won’t be up as much as it’s been up certainly last year. But I mean, you’re thinking about it in the right way. And one of the things that we’re fortunate enough to do is if you think about what we would — the typical company would do if you didn’t have the skill sets and how it should have to do a layoff, that’s expensive and then you have to go hire new talent.

We’re fortunate enough in the areas that Tom talked about, whether it’s development or build out of the network, the sales functions that we can shift those resources over. So if you think about the cost to build out what we’re doing, would cause somebody else a significant amount of money, but we’re able to shift those resources. And as far as the new headcount going forward, we’re going to be pretty judicious with where we’re hiring, but it’s going to be primarily in the areas of cloud and security and maybe a few on the go-to-market side.

Keith Weiss

Got it. And the shift take place both in terms of development as well as go-to-market? Or is it more so like the development resources are being shifted?

Tom Leighton

There’s a lot of resources being shifted within our platform and delivery organizations. Think of the hundreds of people that are managing our network deployment that build out our 4,100 regions today and the tremendous scale we have, and they are now heavily focused on building out compute resources. Think of the people that manage the operations, the automatic deployment of software, the payload or the load balancing, the resiliency, they are heavily focused on incorporating those capabilities for compute, building on top of the Linode framework. So I would say the large majority of the resources are of that nature that we’re retasking.

Operator

Our next question will come from Tim Horan with Oppenheimer.

Tim Horan

Also on the cloud, can you talk a little bit about your improvement in price and performance versus the comments? Can you give us some metrics there? And then can you talk a little bit about what type of cloud you’re building out for? I know, Tom, in the past, you were a little skeptical that we could do kind of gaming as a service over cloud infrastructure. Is that changing? And you could optimize your cloud for AI or blockchain or more networking or very, very low latency? Just any color there? And then lastly, have you contemplated maybe partnering with some of the larger cloud providers with AI really starting to take off? Can you be a partner to a few of them or one of them that would really accelerate things?

Tom Leighton

Yes, good questions. We’ll be the most distributed cloud services provider. And of course, we start with 4,100 POPs for function as a service, JavaScript at the edge. And we’re building out the core cloud capabilities. And then this notion of the — of a more distributed containers closer to the edge, VMs closer to the edge. Now that will give you better performance for things where you want to be close to the end user because it will be closer to more end users will be in some cases, countries where you don’t have a presence from the hyperscalers. And our pricing will be less already. You can look at the list pricing and see that it’s less than what the hyperscalers charge. And we’re — because we’re integrating it with our backbone and with our edge platform, that gives us great economics on the delivery, taking the data in and out of storage or compute and getting it to end users. We’re in a position to do that at a lower cost and to give consumers a better price point.

Yes, this works not for all gaming functions, but for a lot of the things you want to do with gaming. It’s a great use case, streaming, obviously, transcoding, APIs, chatting APIs, people communicating during a sporting event. That’s all — is very relevant to having a more distributed model. AI, I think elementary stuff you can put in a container or VM yes, that makes perfect sense. If you want to have the monolithic storage associated with that, that’s more cloud, core cloud compute, I would say.

I think in terms of partnering, yes, we have a lot of customers that obviously use Akamai services today as well as the cloud giants. In fact, the cloud giants themselves, a couple of them are very large Akamai customers. And so they also use their own services. So I think it’s an ecosystem where, yes, I think there’ll be customers that would use us and use the hyperscalers depending on the application and what they’re looking to do. And we very much believe in a multi-cloud approach.

Operator

Our next question will come from Mark Murphy with JPMorgan.

Mark Murphy

Is it possible to ballpark the revenue contribution that was driven by this huge scale of the World Cup so that we could then remove that from our models for the next 3 years and then, I guess, presumably included back in year four?

Ed McGowan

Yes. Mark, it’s Ed. There’s about $5 million, roughly.

Mark Murphy

Okay. Got it. And then as a follow-up, just to clarify, during Q4, were you able to capture some of the business that Amazon is losing either due to the lower pricing structure that you have for Linode or because you have this advantage of broader points of presence? I guess I’m just interested in — it sounds like from what you’re describing, that potential is there. I’m just wondering if there was a material benefit or tailwind from that in Q4?

Tom Leighton

Not in compute in Q4. Obviously, we compete very successfully in delivery and security with the hyperscalers. We’re the market leader there. Now in compute, they are the market leaders by far. And so nothing that we would do in compute is going to make any difference to them. I mean we’re looking to get to a 1% market share in a market that’s $100 billion to $200 billion a year.

So it will take us a while in compute before I think a hyperscaler would even really, really notice. And of course, a 1% or 2% market share means a lot to us in compute, that’s several billion dollars means less, obviously, to folks at the scale of those guys.

Operator

Our next question will come from Rishi Jaluria with RBC.

Rishi Jaluria

Wonderful. I’ve got two. First, I wanted to start on the compute side of the equation. I appreciate all the color around CapEx. And please forgive me if my math is shouty over here. But if I just do a rough back of the envelope numbers, right, with 9% of total revenue being CapEx, specifically for the compute business and then I strip out $100 million of nonrecurring, that’s still telling me that CapEx — compute CapEx this year in 2023 is going to be 45% of compute revenue, which, again, it’s a growing business and everything. Number one, am I directionally thinking about this? And maybe number — piece number 2 to that is, how should we be thinking about steady state CapEx for the compute business once we get through maybe the next year or 1.5 years? And then I’ve got a follow-up.

Ed McGowan

Yes, sure. So you’re doing the math right. So the way to think about the $100 million, that will enable us for our internal use, that will enable us to save lot more than $100 million. So think of it — you think about it right in terms of as a onetime sort of burst of a charge. Obviously, if we continue to use our own compute capabilities over time down the road. We will be adding a little bit from time to time there. But you’re thinking about that correctly.

In terms of what does the steady state look like? We talked at the Analyst Day about how you can approximate sort of future growth percentage with what you would spend. So for example, if we are — of the $1 billion we’re spending 30% in CapEx, our growth rate would be probably around 30%. It’s not quite dollar for dollar, but it’s sort of a rough approximation. So think of us spending roughly $100 million this year on internal use another, say, call it, 225 to 250 depending on where you are on your models. That enables you to get that type of revenue scale potentially. Obviously, it’s going to have to play out in the market to be a little bit more if we get a little bit less. That’s a pretty decent rule of thumb.

And then obviously, as we’re growing out these locations, we’re pretty ambitious in terms of the core locations that we’re putting online also distributed locations are putting online. We’re going to be investing ahead of revenue. So I would expect for the next 1.5 years or 2 years to have elevated CapEx related to the compute business, and we’ll start to scale into it.

Rishi Jaluria

Okay. Got it. That’s helpful. And then just maybe going back to the security business. Going past this year, longer term, what needs to happen to see security overall reaccelerate growth rate that you’d be happy with? Is that primarily going to be driven by a continued Guardicore mix shift? Is that going to be more on the Zero Trust portfolio ex Guardicore? Maybe walk us through kind of what needs to happen to get security up to kind of an organic growth rate that you’d be happy with?

Tom Leighton

Yes. No, I think you characterized it well, and it is more the mix shift to the newer products for us that are growing at a rapid rate, but are still relatively small. We’ve got a substantial return from Bot Manager now. That’s a great example, starting to really help. Next is Guardicore, which, as Ed mentioned, we want to — as we exit should be at over $100 million run rate. And once you start getting to that size and it’s rapidly growing, it starts making a difference for the big security number. And of course, as that number gets bigger, it obviously gets more challenging to maintain the higher growth rates. We’re continuing to invest in new capabilities there. Account Protector is off to a very good start, really excited about that. But it will take time to do that. And we’re continuing to look for potential acquisitions that can help jump start growth. I don’t think anything huge. It gets you the return right away, but areas that we think are really important that we can become market leaders in like we’ve done for application firewall, bot management, for segmentation and that over then a period of years can drive significant growth for us.

Operator

Next question will come from Amit Daryanani with Evercore.

Amit Daryanani

Two questions for me as well. I guess the first one, when we think about 2% to 4% kind of top line growth in ’23 in constant currency, I think you talked about security growing double digits. Is there a way to think about how do you think growth stacks up on the compute and delivery side as well for the year?

Ed McGowan

Yes, sure. So well, we didn’t give specific guidance for either delivery or compute. Tom did talk about us achieving $0.5 billion or so in compute revenue. So depending on where you put your models, you decide the $0.5 billion for compute and just solve for the delivery. I think if you were to be on the higher end of the business, delivery might do a little bit better as we saw in Q4, potentially get security going. If the macroeconomic conditions improve and then obviously there really is just a timing issue in terms of when we can start moving major workloads on compute.

Amit Daryanani

Got it. And then as we think about sort of the path from, let’s just say, 27.5% operating margins that you’ll be at in ’23, towards just 30% kind of target you folks have had in the — you have now actually, what do you think takes to achieve that target? Is there a revenue number that you need to get there or a mix or the cost reduction? If you could just maybe provide a bit of a bridge on how do you get from 27.5% to 30% EBIT margins, that would be really helpful?

Ed McGowan

Yes. So it’s a little bit of everything, right? But I think as we laid out about five or six different things that we’re doing, probably the biggest near-term item would be the third-party cloud costs. So as we migrate that our own cloud, we’re going to save about $100 million or so. Think about that as some of that will happen this year, a lot more of it in ’24 and then by ’25, we should have almost the majority of our internal cloud or third-party cloud on our eternal system. So that’s 2 points to 3 points right there.

Colocation, I talked a bit about the audit of the ASC 842 in lease accounting. We have a little bit more of a burden that we have to take at the beginning of some of these longer-term agreements. But also, we’re spending on colocation that we haven’t quite scaled into yet. So there’s your question about revenue. As revenue scales, you’ll get scale with your margins. And then just through some of the depreciation savings that we’re getting on delivery and the real estate savings. We’re spending, call it, before this year around $100 million in real estate, we’re going to save about $20 million this year. Probably room to get maybe another $20 million or $30 million out of that as well. So it’s really a combination of getting the compute revenue to scale into our investments in mostly our colocation facilities and third-party cloud savings, those are probably the two biggest areas for margin expansion.

Amit Daryanani

Perfect. It sounds like most of the ramp to margins is going to be driven by self-help levers versus revenue tailwinds. I mean, it seems like the skew might be a bit more on self-help. Is that a fair way to characterize it?

Ed McGowan

Yes. I mean I think that’s — when you look at the size of those numbers, certainly, I mean, obviously, revenue cures all your rails, right? We’ve got a pretty scalable model. So if we see acceleration in revenue, especially on the compute side, you’re going to see a pretty good flow-through. But yes, there’s a lot in our control here. And I think we’re doing the responsible things as far as hiring goes, shifting resources, focusing on reducing our real estate costs and really focusing on driving down that third-party cloud cost, which is really taken a pretty big chunk out of our gross margin.

Operator

Our next question will come from Fatima Boolani with Citi.

Fatima Boolani

Two for me. On the delivery segment, the last couple of years for you have been maybe a little bit more erratic just by way of lapping some of the benefits from the pandemic, some of the dynamics you saw in the gaming area where trends are moderating. So at a high level for you, when we think about the underlying cadence from a volume perspective, what are you assuming that’s maybe different the same, better or worse versus the trends you saw this year? And I’m considering the renewal cohort what you’ve mentioned around holding on pricing. So any sort of color commentary you can give us for the delivery segment in terms of traffic trends and then pricing trends under the hood? And then I have a follow-up, please.

Ed McGowan

Yes, sure. Great question. So let me see if I can pick that all at once here. So in terms of the major renewals, we don’t have nearly as many major renewals as we had last year. So that’s going to be one thing that works in our favor. We are not anticipating in our guidance any significant increase in traffic in terms of levels of growth that we saw last year. We’re anticipating a slight increase, but nothing significant. We’re anticipating that and we’re starting to see that price declines are moderating a bit. They’re not nearly as steep as they’ve been in the past. And then we will continue our posture in terms of being a little bit more selective with some of the spiky traffic. Once we start seeing volumes get back to the historical Internet growth rates and beyond, then maybe that posture may change, but that’s the way we’re going to play it for now.

Fatima Boolani

I appreciate that. And just the delineation between your U.S. business and the international business, anything you can point to by way of geographical differences in procurement? Is there more sensitivity on budgets in the U.S. versus international? Any characterization there because international continues to be a stronghold for you versus some of the — maybe some of the malaise on the U.S. side, but would love to get a little bit more granular on what’s continuing to drive that strength and if some of that budgetary pressure that you alluded to on the security side is maybe showing up more pronouncedly in U.S. versus international? That’s it for me.

Ed McGowan

Sure. Yes, I would say just kind of general macro across all regions. I’d say new customer acquisition is more challenging in an environment like this. And I think you hear a lot of companies talk about that. We’re seeing that as well.

In terms of geographic, obviously, the European economy is struggling a bit more than we are in the U.S., slightly higher inflation, et cetera. So I expect that area to be a bit more weaker than what I’m seeing in the U.S. Asia is still been pretty strong. Latin America has been pretty strong. U.S. has been kind of holding firm here. But I would expect the European business, in particular, is to be the most impacted by the macroeconomic factors.

Operator

Our next question will come from Frank Louthan with Raymond James.

Frank Louthan

So with the significant number of POPs that you have already, what is it about the cloud platform that you need to expand these sites? What is about the location and the capabilities that they bring for expanding the compute platform that you generally have? And then you touched on a possibility for M&A. What do you consider significant? And what are the — what size M&A do you think you might be looking at if you do any in the next 12 months?

Tom Leighton

Yes. So on the Akamai Connected Cloud architecture, we already have 4,100 edge regions. And these regions do delivery and security. They’re the first line of defense. They also do what we call edge computing, which is function as a service, JavaScript at the edge. Now in the core, where Linode had 11 data centers, and we’re going to more than double that, you have very large-scale storage object and block storage. You’ve got VM as a service, Container as a Service, Kubernetes, core cloud compute.

Now we’re adding also an intermediate layer, we call those sort of the distributed compute layer. And this would have not the monolithic storage, but you’d have containers as a service. Kubernetes, VMs as a service, so you could do compute there. And so what you do and where you want to do it depends on the application, things that are a lot of back and forth with the end user that are lighter weight, that could be handled in JavaScript. You want to be doing that in the 4,100 edge regions, something like transcoding, you’re doing some data processing, you got an app in a container, but performance matters, you want to be close to the end user or say a gaming application that you’ll want to do in our distributed edge platform.

And the reason we have that is there’s a lot of cities and places in the world that don’t have a giant cloud data center there. And they can’t really take advantage of the cloud for applications that have — where the proximity to the user is important, and so that’s something that we want to be able to provide. So there’s three sort of levels here of compute and you want to be actually using generally all three for a major company, but the specific application dictates where you want to be doing it and what combination that you want to use.

Frank Louthan

Okay. Potential for M&A and thoughts on what you would consider significant versus more tuck-in?

Tom Leighton

Yes, good. So I think tuck-ins are a team that has the beginnings of a product or technology that we can scale on behalf of our customers. We’ve done a lot of acquisitions like that. More substantial would be something like Guardicore that had, at that time, the #2 product in the marketplace, and we’ve invested around that. They’re now #1, which is great to see and a more significant cost associated with that. And they had a developed product already. Go back farther Prolexic is another example of that where they had a developed product, they already had $40 million, $50 million in ARR. And those are — we do occasionally, and we’re always looking. But we do those occasionally. But more you’ll see the tech tuck-ins and then we develop from there.

Operator

Our next question will come from Rudy Kessinger with D.A. Davidson.

Rudy Kessinger

I guess I’m curious, how did the macro trend in the quarter just with deals lengthening? And how many deals did you see push, I guess, relative to Q3? And the guide assumes no positive or negative change in the macro. Just why make that assumption, why not be a bit more conservative, maybe assuming it gets a little worse?

Ed McGowan

Yes. So this is Ed. So in terms of the biggest impact on the quarter, I would say coming in, we’re expecting potentially a weak commerce season. We actually saw a pretty decent commerce season. We saw a decent video traffic. We saw obviously, we talked about the World Cup being better than our expectations. I would say gaming was noticeably weak. We didn’t see as much activities you typically see.

And then from the sales side, a couple of deals pushed a few large Guardicore deals we track that push and then also new customer acquisitions are slower than we’d like to see. Those are really the big things. Then as far as the macroeconomic conditions, it’s tough when you sit in the seat to try to play economist. And with us, obviously, exit has probably the biggest impact on us more than anything, given we’ve got longer-term contracts.

But it’s hard to say that, look, things can change and get dramatically worse. I’m just expecting what I see right now in front of us in terms of the macroeconomic environment is challenging, but I think we can navigate it pretty well. So that’s what we based our guidance on. If things change, we’ll obviously update the guidance. But I was just trying to — a lot of times people asked me what was what we were thinking as you put your guidance that are you expecting things to get dramatically worse. In this case, no, I expect things to be dramatically better not kind of roughly the end if things do change, we’ll obviously update you as we get more information.

Rudy Kessinger

Yes. Okay. That’s fair. And then at a higher level, I mean, obviously, it sounds like compute has kind of slotted into growth priority one, if you will, kind of ahead of security. And just at the high level, I’m curious what really is giving you conviction to invest so much in compute.

Is it the early customers that you’ve signed? Is it just conversations? Is it the pipeline growth you’ve seen over the last few quarters since you made the acquisition? What’s giving you so much conviction to make such a large investment go-on on compute?

Tom Leighton

Well, it’s all of the above plus the compelling logic of the situation. As I mentioned, I do have the opportunity to meet with the senior most executives that a lot of the world’s largest companies, especially you think media, commerce, gaming and so forth. And there is a lot of interest there in our ability to help them. They’re in a situation where they’re spending a ton on third-party cloud. It’s growing rapidly. Many of them describe it as being out of control. and it’s even hard to know how big it will get.

In some cases, they’re spending this with a competitor. And on top of it, they’ve got major company initiatives to cut cost because of the challenging global economic conditions. And so when we can offer them a service with at least as good, better performance, at a lower price point, that’s very attractive.

Now on top of it, these companies, they know us well. They already trust us with scale and performance and security because we provide the vast majority of their delivery and security. So we are a very logical choice. It’s not like we’re just somebody coming along here saying, hey, we got a cloud service. It’s not like that at all. We’ve got a lot of credibility with these companies and they are pretty clear that they think this could be very attractive for them. In fact, I think one of the analysts on this call did a survey of 50 of our larger customers have found the same thing. The same thing was reported to them.

So we see that our customers need that and would like to shift business to us, but we want to get ready for that and help them. And that means building out the capacity building out the new distributed architecture and getting the functionality to the level that they’re going to be comfortable putting mission-critical applications at very large scale on our platform. And of course, Akamai will be one of the first examples of that. We’re going to place our services that are used by pretty much all the major enterprises, a lot of the major enterprises out there for, for example, security onto our platform. And that will be another great proof point that Akamai Connected Cloud is really going to work for them.

Operator

Our next question will come from Ray McDonough with Guggenheim.

Ray McDonough

Tom, maybe just to ask the M&A question in a slightly different way. As we think about the strategic plan going forward, do you feel the organization has enough operational capacity to continue to expand Linode right now and do another acquisition in security at this point to help you reaccelerate growth? And is 20% long-term growth including acquisitions in the security business, something you are still targeting after this year?

Tom Leighton

Yes, good question. First, the work on Linode which is extensive is use different teams by and large than our security technology group. And so yes, we are in a position that we can continue that work and also do security acquisition. I don’t think there’s any challenge there.

Now we have, as we talked about, retask a lot of the positions, either people or positions from our platform and delivery organizations to the compute effort. So there is a lot of effort there, and that will be increasing throughout the year.

Now in terms of the 20%, what we’re talking about is we’re guiding this year into the low double digits and then we’ll see where we are. Obviously, we’d like to be growing faster than that, but there are very challenging conditions out there, as we’ve talked about, and we’ll have to see. So we’ll update that guidance as we — maybe we get to an Investor Day or get into next year. But right now, we’re just guiding for this year in security in the low double digits.

Ray McDonough

Great. And one more, if I could. The work we’ve done on Land does suggest your customer base is interested in leveraging Linode. And when I think about sort of the lead times for capacity additions in the data center space that you’ve already leased, how long are the lead times to fill that data center space? And at what point would you expect it at least to get to a somewhat full utilization of just the space that you’ve already contracted for?

Tom Leighton

Yes. So good question. There’s the contracting, there’s the build-out, there’s getting it all turned on. And the way you want to think of it is we would be doing the core data center build-out. The focus of that is in the first half, and we’ll have a lot of that done then, and we should finish that early in the second half or get substantially more than we have today.

And then late in the second half, taking on the distributed architecture, we should have a bunch of those regions turned on live later in the year. And at that point and also, at the same time, getting the certifications in place, virtual private cloud, other capabilities that the big enterprises need so that when we get to the later part of the year, we’re in a position that we can take on this business, and it will just happen all at once. A major enterprise will try it, use some applications and then some of the big ones, you port some of the traffic over.

And so I think the major feeling of it and use of it really comes more in ’24, not so much this year. We expect, as we talked about this year to do about $0.5 billion in compute. But the real growth, I think and the monetization of what we are building out now comes in ’24 and ’25. And also for our own use. This year, we are in the process of porting our own applications. And as Ed talked about we are going to save some this year but really the big savings for us comes in ’24. So that’s the way I think to think about it.

Tom Barth

Well, it is Valentine’s Day. So we’re in a little bit over. But operator, why don’t we take one more question? I know there’s probably a few more, but let’s just take one more and we’ll end the call.

Operator

Okay. Our last question will come from Michael Elias with Cowen & Company.

Michael Elias

Just two questions for you. The first is you talked about some Guardicore deals slipping. And last quarter, there was commentary around longating sales cycles. Just wondering if you could give us some color on how sales cycles have evolved for not only Guardicore but the broader portfolio? And then my second question for you is maybe to phrase it a different or frame it a different way, is there a rule of thumb that we should use to think about the correlation between incremental revenue and incremental megawatts of data center capacity that you need, i.e., to support an incremental $100 million in compute revenue, you would need x amount of megawatts?

Ed McGowan

All right. I’ll try, the second one is going to be a little bit more challenging and maybe I’ll come back with a metric on that one. I don’t have exact metric down to the megawatt, but I’m sure somewhere in the business, I can find somebody who does. But I’ve sort of used sort of a rule of thumb of $1 of CapEx is $1 of revenue roughly speaking, that may change a bit, but sort of a good rule of thumb.

In terms of sales cycles, the good news with the deals slipping is they’re not going away. This is typically a big purchase that you make, especially with Guardicore that it could be something where the decision-maker is just pushing it off a quarter or two. And it’s not that the deals are losing. It just becomes a longer sales cycle and a fight for budget.

In terms of just overall sales cycles, like I said, the biggest impact is new customers, right? It’s a lot easier with your existing customers, you’re going through renewal cycles, you’re having upgrade conversations and things like that, getting new customers to open up a new buying pattern is challenging. I do think one thing that will work in our advantage in all of this, though, is as we talk about compute and the need to find a cheaper alternative go multi-cloud. I think that’s going to work in our favor.

So I think one thing this — all the negatives that come with the macroeconomic backdrop that we have. I think that’s one positive that will work in our favor, and we are starting to see our pipeline grow pretty dramatically in that area. And as Tom talked about, we should start to see deals close towards the back half of the year and then really set ourselves up for ’24.

Tom Barth

Okay. Well, thank you, Michael. Thank you, everyone. And in closing, we’ll be presenting at several investor conferences and events throughout the rest of the first quarter. Details of these can be found in the Investor Relations section at akamai.com. We want to thank all of you for joining us, and we wish you a very good health and good health to your family. So have a nice evening. Take care.

Operator

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