Cloudflare (NET) Q4 2024 Earnings Call Transcript

admin6 hours ago

NET earnings call for the period ending December 31, 2024.

Image source: The Motley Fool.

Cloudflare (NET -1.92%)
Q4 2024 Earnings Call
Feb 06, 2025, 5:00 p.m. ET

Prepared Remarks:

Operator

Good day, everyone, and welcome to the Cloudflare Q4 2024 earnings call. This conference is being recorded. At this time, I would like to hand the call over to Mr. Phil Winslow.

Please go ahead, sir.

Phil Winslow — Vice President, Strategic Finance, Treasury, and Investor Relations

Thank you for joining us today to discuss Cloudflare’s financial results for the fourth quarter of 2024. With me on the call, we have Matthew Prince, co-founder and CEO; Michelle Zatlyn, co-founder and president; and Thomas Seifert, CFO. By now, everyone should have access to our earnings announcement. This announcement as well as our supplemental financial information may be found on our Investor Relations website.

As a reminder, we will be making forward-looking statements during today’s discussion, including, but not limited to, our customers, vendors, and partners’ operations and future financial performance; anticipated product launches; and the timing and market potential of those products. Our anticipated future financial and operating performance and our expectations regarding future macroeconomic conditions. These statements and other comments are not guarantees of future performance and are subject to risks and uncertainty, much of which is beyond our control. Our actual results may differ significantly from those projected or suggested in any of our forward-looking statements.

These forward-looking statements apply as of today, and you should not rely on them as representing our views in the future. We undertake no obligation to update these statements after this call. For a more complete discussion of the risks and uncertainties that could impact our future operating results and financial condition, please see our filings with the SEC as well as in today’s earnings press release. Unless otherwise noted, all numbers we talk about today other than revenue will be on an adjusted non-GAAP basis.

You may find a reconciliation of GAAP to non-GAAP financial measures that are included in our earnings release on our Investor Relations website. For historical periods, a GAAP to non-GAAP reconciliation can be found in the supplemental financial information referenced a few months ago. We would also like to inform you that we will be participating in the Morgan Stanley Technology, Media, and Telecom Conference on Monday, March 3. And we’re hosting our Annual Investor Day on Wednesday, March 12.

Now, with that said, I’d like to turn the call over to Matthew.

Matthew Prince — Co-Founder and Chief Executive Officer

Thank you, Phil. We had a very strong end of 2024. We achieved revenue of $459.9 million, up 27% year over year. During the quarter, we added a record number of new large customers, those that pay us more than $100,000 per year, and now have 3,497 large customers, also up 27% year over year.

Revenue contribution from large customers grew to 69% of revenue, up from 66% in the fourth quarter last year. Our dollar-based net retention ticked up one percentage point quarter over quarter to 111%. Our gross margin was 77.6%, remaining above our long-term target range of 75% to 77%. We delivered an operating profit of $67.2 million, representing an operating margin of 14.6%.

We continue to generate strong free cash flow, achieving $47.8 million during the quarter and $166.9 million for the full year. As we’ve talked about multiple times, since the beginning of 2024, customers have been disciplined with their budgets, scrutinizing deals carefully and ensuring every dollar spent delivered clear and immediate value. That trend continued through Q4. However, as the quarter progressed, we saw encouraging signs that confidence is beginning to return, particularly in the U.S.

Security, AI, modernization, and efficiency form the word cloud we hear most often in these conversations. These themes play directly to Cloudflare’s strength. Beyond the qualitative, we saw measurable improvements in Q4. We saw a notable uptick in close rates.

We saw an improvement in sales cycles. The majority of large customers I mentioned last quarter, those deals had slipped from Q3, reengaged, and signed significant contracts in Q4. We crossed 3 million active developers on our platform, including Cloudflare Workers and Workers AI. And we saw record growth in our largest customers, those that spent over $1 million with Cloudflare per year.

We ended the year with 173 such customers, 55 of which we added in 2024, and more than half of the new adds were in the fourth quarter alone. I’m proud of how our team remains disciplined and focused on delivering real ROI for customers, which drove these record results in the fourth quarter, while also ensuring we’re well-positioned to capture the demand we see lined up in 2025. More than anything, behind this success is our improved go-to-market execution. I wanted to give you an update on our progress on that front.

Mark Anderson continues to prove he’s one of the best go-to-market leaders in the industry. In the fourth quarter, he led the team to a fifth consecutive quarter of double-digit year-over-year increases in sales productivity. For the full year, not only with the average productivity higher each quarter compared with 2023, but we also achieved meaningful improvements in shifting account executives to the right of their attainment graph. We delivered a 10-percentage-point increase in ramped AEs, achieving over 80% of quota compared with 2023, with most gains coming in the 125% or higher attainment cohorts.

We continue to aggressively hire in our sales organization with a focus on more stage-appropriate talent and onboarding enterprise account executives with proven track records. In Q4, nearly 80% of our new sales hires were in the enterprise segment and the absolute number of new enterprise AEs hired increased 84% year over year. You’ve heard me say in the past, we’re not limited by our total addressable market by our competitors or by our pipeline. Each day, I see evidence this is truer than ever before.

Our constraint has been and at least for another quarter, will continue to be the capacity of our sales force. Now, we’ve seen the increase in our sales force’s performance under Mark’s leadership. Net sales capacity turned the corner exiting 2024. In Q2 of 2025, we’ll start to see capacity and ramp reps begin to meaningfully accelerate.

To give you a sense, 80% of our full-year plan for 2025 is assigned to account execs who are already in their seats at Cloudflare at the start of the year. And now that we’ve got the formula dialed in, we’re keeping our foot on the gas. This is what gives me confidence in our ability to reaccelerate throughout this year. We are world-class in product innovation.

We are world-class in network stability and reliability. And 2025 is the year we will prove we can be world-class in go-to-market as well. That seems like a great segue to discuss some of our customer wins in the quarter. A Fortune 100 technology company signed a five-year $20 million pool of funds contract which includes all Cloudflare products for frictionless adoption across our entire platform, marking the largest new customer win in Cloudflare’s history.

Expected initial use cases include application security and performance as well as our Workers developer platform. They view Cloudflare as a strategic partner with a shared vision for multi-cloud security, AI, and data sovereignty. This is just the beginning with this customer. A leading AI company expanded their relationship with Cloudflare, signing a 1-year $13.5 million pool of funds contract.

Deploying this strategic pool of fund structure will provide seamless access to the entire Cloudflare platform, maximizing scale and enabling accelerated innovation with transparent pricing and a streamlined partnership. In the words of the customer, “Cloudflare is a model partner. We wish to emulate this relationship with all our vendors.” A leading global retailer signed a three-year $10.8 million contract for application services, Workers, R2, Magic Firewall, and Magic Transit. This customer was looking for a strategic partner who could help further develop and scale its online experience, which has been experiencing significant growth.

Displacing a 20-year incumbent, Cloudflare’s unified offering and vast global presence, combined with our innovative developer platform, including Workers AI, proved to be key differentiators from the incumbent and other competitors. A major U.S. investment firm expanded their relationship with Cloudflare, signing a three-year $4 million SASE contract for Zero Trust and data loss prevention, along with Magic WAN and Magic Firewall. This customer approached Cloudflare unhappy with their incumbent solution from a first-generation Zero Trust provider.

Cloudflare won against multiple competitors due to our superior network performance and ability to deliver a transformational single-vendor SASE solution on an easy-to-use unified platform. A Global 2000 aviation group signed a five-year $9.4 million contract for our full suite of application security and performance products. This customer is focused on adopting a cloud-first operating model to increase agility, while also improving cost efficiency. The company conducted a robust RFI against more than a dozen vendors in consideration.

And Cloudflare was selected to displace a longtime incumbent due to our ability to drive greater modernization, the ease of use and flexibility of our products, and our ability to deliver meaningful ROI. A Global 2000 financial institution signed a four-year $13.6 million contract for application services along with Magic Transit and Threat Intelligence. This customer is looking to accelerate digital transformation as well as gain more control and transparency over its security posture. With their nine-year incumbent provider, it would take 14 weeks to provision a new website.

With Cloudflare, a new website can be provisioned in just 10 minutes. In the words of this customer, “Cloudflare stands out in this space, offering a robust, scalable, and developer-friendly Edge Security platform that not just offers best-in-class security for our digital channels but also helps us increase velocity of digital changes to speed up the bank.” A rapidly growing technology company expanded their relationship with Cloudflare, signing a two-year $1 million contract for Cloudflare Calls. Cloudflare Calls is a newer product in our developer platform, allowing developers to build real-time audio and video applications on Cloudflare’s serverless Edge. This customer was looking for a more cost-effective and performance solution.

With Cloudflare, this customer was able to build new custom use cases that can scale significantly while also achieving better performance at a lower price point. Finally, a major international financial institution expanded their relationship with Cloudflare, signing a three-year $6.1 million contract for application services. This customer is migrating from a hyperscaler to Cloudflare in order to enable a multi-cloud architecture as they scale their growth in a regulated industry around the world. Cloudflare’s best-of-breed security products, higher uptime, and ease of use to scale and automate operations made us the clear winner.

I wanted to end by talking a little bit about what we’re seeing in AI. We believe Cloudflare has four distinct opportunities in AI. The first is the same as every other company. We’re getting more efficient in our business processes using AI systems.

That’s not particularly interesting these days. Though I’m proud, more often than not, we’re finding we can build these functions on our own infrastructure rather than contracting with others. The second is AI makes our performance and security products smarter for customers. At some level, although we never really had the hubris to describe it this way, Cloudflare has always been an AI company.

The thesis was that if we could get enough Internet traffic flowing through us, we can spot security threats that no one else could see. And today, our machine learning-based security systems regularly discover new security threats that no human had identified before. That, again, feels like table stakes for us. The third opportunity is where I think things start to get interesting.

The killer application for Cloudflare Workers is turning out to be AI. The model of programming is uniquely suited for building tools like AI agents, and our serverless architecture, which allows you to pay only for what you use based on CPU or GPU type, positions Workers to become the go-to platform for developers who want the best price performance for AI inference and agentic workflows. I talked last quarter about a large AI customer that was building their interface on top of our inference platform. From their perspective, the partnership has gone extremely well.

But behind the scenes, our team has been able to do the hard engineering work to drive the efficient use of our GPU infrastructure. Inference tasks are generally highly variable, and there’s a two times difference between these customers’ peaks and valleys, which means they would have to pay approximately 250% more to stay provisioned on a hyperscaler to run the same number of inference tasks compared with Workers AI’s efficient pay-per inference serverless model. Additionally, more and more developers are discovering AI Gateway, with some realizing more than 10x price performance improvement for their AI agents by serving requests directly from Cloudflare’s cache instead of the original model provider. And just last month, the world was amazed that the efficiencies of the team of clever engineers in China were able to deliver in the field of AI training with the DeepSeek model.

We are seeing that there are equivalent optimizations that can be made with AI inference on Cloudflare’s platform, resulting in faster performance and lower prices for customers and higher margin, and less capex for us. We believe inference is a bigger opportunity than training and our team continues to find step-function breakthroughs that put us well ahead of any alternative. But all that may pale in comparison to the fourth opportunity. Cloudflare counts many of the most important AI companies as customers.

We also count a huge portion of the world’s content creators as our users. Being between those two puts us in an important role to help figure out the business model of the post-search web. Cloudflare sits in a unique position to help figure out how content creators are compensated, what agents are allowed where and on what terms, and how the AI-driven web of the future will fit together. It’s early days, but the conversations we’re having with all the relevant parties feel foundational for the future.

Watch this space, definitely exciting times. But before I get too far ahead of myself, to bring it back to the present, let me hand it off to Thomas to talk through the fourth quarter’s financials and our outlook for 2025. Thomas, take it away.

Thomas J. Seifert — Chief Financial Officer

Thank you, Matthew, and thank you to everyone for joining us. We are pleased with our strong operational and financial performance for the fourth quarter, building further momentum from our go-to-market transformation efforts that Matthew discussed. Strengths in our business this quarter was driven by significant growth with large $1 million customers, ongoing traction with pool of funds contracts, sustained momentum with our Workers developer platform, and high prioritization of security by our customers. We saw a notable uptick in close rates and an improvement in sales cycles.

We also delivered another double-digit year-over-year improvement in sales productivity with record productivity in both EMEA and APAC. As discussed last quarter, we were encouraged to see the number of ramped account executives increase exiting the fourth quarter, and we expect year-over-year growth in ramped AEs to continue to accelerate each quarter throughout 2025, further laying the foundation for Cloudflare’s next phase of growth at scale. Turning to revenue. Total revenue for the fourth quarter increased 27% year over year to $459.9 million.

From a geographic perspective, the U.S. represented 50% of revenue and increased 23% year over year. EMEA represented 28% of revenue and increased 27% year over year. APAC represented 14% of revenue and increased 39% year over year.

Turning to our customer metrics. In the fourth quarter, we had approximately 237,700 paying customers, representing a record addition of nearly 48,000 paying customers in 2024 and an increase of 25% year over year. We ended the quarter with about 3,500 large customers, representing a record addition of more than 740 large customers in 2024, an increase of 27% year over year as well as a record addition of 232 large customers in the fourth quarter alone. We were pleased to see revenue contribution from large customers during the quarter increased again to 69% of revenue, up from 66% in the fourth quarter last year.

For full-year 2024, revenue from large customers represented 67% of total revenue compared to 64% in 2023 and 61% in 2022. And we ended the year with 173 customers that spent over $1 million with us, adding a record 55 $1 million customers in 2024 and representing a 47% increase year over year. Our dollar-based net retention was 111% during the fourth quarter, representing an increase of one percentage point sequentially. As a reminder, there can be some variability in this metric quarter to quarter, but we believe the recent decelerating trend in DNR is stabilizing despite continued near-term headwinds from increased traction with pool of fund contracts, which can impact the shape of revenue recognition, especially for existing customers transitioning to these platform deals.

Moving to gross margin. Fourth quarter gross margin was 77.6%, representing a decrease of 120 basis points sequentially and a decrease of 130 basis points year over year. During the fourth quarter, the percentage paid versus free customer traffic increased as compared to its prior quarters, resulting in a higher allocation of expenses to cost of goods sold from sales and marketing. The underlying economics of our network driven by its inherent scalability and efficiency remained unchanged.

Network capex represented 50% of revenue in the fourth quarter and 10% of revenue for the full year. As we mentioned last quarter, the accelerating shift from AI training to AI inference has given us confidence to continue to increase our investment in our GPU rollout as we provision greater capacity to support demand in 2025. As a result, we expect network capex to be 12% to 13% of revenue for full year 2025. Turning to operating expenses.

Fourth quarter operating expenses as a percentage of revenue decreased by 5% year over year to 63%, as we remain committed to driving higher productivity and greater efficiency across our operations. Our total number of employees increased 16% year over year, bringing our total head count to about 4,300 at the end of the quarter. Sales and marketing expenses were $166.9 million for the quarter. Sales and marketing as a percentage of revenue decreased to 36% from 40% in the same quarter last year.

Research and development expenses were $74.8 million in the quarter. R&D as a percentage of revenue remained consistent at 16% compared to the same quarter last year. General and administrative expenses were $47.8 million for the quarter. G&A as a percentage of revenue decreased to 10% from 11% in the same quarter last year.

Operating income was $67.2 million, an increase of 69% year over year compared to $39.8 million in the same period last year. Fourth quarter operating margin was 14.6%, an increase of 360 basis points year over year. These results highlight our continued focus on becoming more efficient and more productive given that operational excellence is a long-term competitive advantage. Turning to net income in the balance sheet.

Our net income in the quarter was $68.8 million or a diluted net income per share of $0.19. Our non-GAAP effective tax rate in Q4 was 25% compared to guidance of 16% to account for the full-year impact of certain tax elections made during the fourth quarter, which did not increase our cash liabilities in 2024, but expected to mitigate cash liabilities in future years. Excluding the impact of these tax election, our reported diluted net income per share would have been $0.22 for the quarter. Maintaining our strong commitment to being fiscally responsible and acting as good stewards of investors’ capital, we ended the fourth quarter with $1.86 billion in cash, cash equivalents, and available-for-sale securities.

Free cash flow was $47.8 million in the quarter or 10% of revenue compared to $50.7 million or 14% of revenue in the same period last year. Remaining performance obligations, or RPO, came in at $1.687 billion, representing an increase of 12% sequentially and 36% year over year. Current RPO was 70% of total RPO, growing 30% year over year in the fourth quarter versus 29% in the third quarter and 26% in the second quarter, respectively. Moving to guidance for the first quarter and full year 2025.

As a management team, we remain deeply committed to the unit economics of our business and focused on the two fundamental drivers of long-term value creation, growth and profitability. We’ve always taken a disciplined data-driven approach to scaling Cloudflare, balancing investments for future expansions with financial and operational efficiencies to ensure that every dollar we deploy drive significant returns in the form of durable long-term growth and profitability. Over the past two years, we’ve demonstrated our ability to drive operating leverage while transforming our go-to-market operations as we pursue the massive opportunity still ahead of us. Our disciplined approach gives us the flexibility to lean in when we see the right opportunities.

And as we enter 2025, the data we have gives us confidence that now is the time to continue to invest to reaccelerate growth. Importantly, this is based on tangible data, we have real visibility into the factors that drive reaccelerating growth. Improving sales productivity and attainment levels, ramping capacity of account executives already in seats, building momentum with large $1 million-plus customers, growing pipeline, higher win rates, and increasing traction in key areas, including Workers AI and SASE. For the first quarter, we expect revenue in the range of $468 million to $469 million, representing an increase of 24% year over year.

Variable revenue is still a new and growing part of our business model. While we saw strong signals during the fourth quarter reflecting increased usage, we are maintaining a prudent outlook for the first quarter, given more limited historical data on consumption patterns and seasonality. We expect operating income in the range of $54 million to $55 million. We expect an effective tax rate of 21%.

We expect diluted net income per share of $0.16, assuming approximately 362 million shares outstanding. For the full year 2025, we expect revenue in the range of $2.090 billion to $2.094 billion, representing an increase of 25% year over year. We anticipate the weighing of revenue in the second half versus the first half of 2025 to be approximately 40 to 50 basis points higher as compared with the relative mix in 2024. We expect operating income for the full year in the range of $272 million to $276 million, and we expect an effective tax rate of 21% for 2025.

We expect diluted net income per share over that period to be in the range of $0.79 to $0.80, assuming approximately 366 million shares outstanding. In closing, we’re excited about the road ahead and confident that our strategy will drive continued innovation and accelerating growth. As always, we will stay focused on creating significant shareholder value with our commitment to disciplined execution, durable growth, and operational efficiency. We look forward to updating you on our progress in the coming quarters as we accelerate Cloudflare’s next stage of growth at scale.

And with that, I’d like to open it up for questions. Operator, please poll for questions.

Questions & Answers:

Operator

Thank you, sir. (Operator instructions) We’ll go first to Hamza Fodderwala, Morgan Stanley.

Hamza Fodderwala — Analyst

Great. Good evening and thank you for taking my question, and congrats on the strong finish to the year. Matthew, obviously, DeepSeek was a big story so far this year. I’m curious, as we see the rate of improvement and the efficiency of these models, whether you think more and more of these inference workloads start moving to the edge.

Thank you.

Matthew Prince — Co-Founder and Chief Executive Officer

Yeah, Hamza. First of all, let me apologize, I have a bit of a cough, so I’ll try my best not to cough into the phone too much during this. But DeepSeek was — first of all, I think there’s been a lot of noise around DeepSeek because it came out of a — from a Chinese company, I think there’s a lot of concern around national security issues and other things. I think we boiled it down and looked at the underlying research behind it.

And the reality from what we’ve seen is that the researchers at DeepSeek came up with two very, very clever optimizations that significantly reduced the cost of training. And I think that’s really interesting if you sort of separate it away from a lot of the noise. My view is that models over time will commodify and will be largely open rather than proprietary. And that’s something that DeepSeek, I think, is the sort of camel’s nose under the tent.

I think what Meta is doing with Llama is another effort in that area. And that’s good for us. That’s good for Cloudflare because we can have a broad set of models that are running on Cloudflare’s network and available to our customers. I think the second thing that was important was that DeepSeek really demonstrates that there is room for efficiency.

For a while there, I felt like we were the only one saying, “Hey, you can make AI a lot more efficient.” I think some of the folks in Cupertino at Apple were seeing the same thing although are much more secretive about what they’re seeing. But this idea that the only way to win in AI was to spend hundreds of billions or even trillions of dollars, I think DeepSeek shook that foundation. And I think that plays very much to Cloudflare’s strength. We are extremely good at wringing out as much efficiency as possible.

And we’re seeing the same opportunities with inference that DeepSeek saw with training, and I think that that gives us a lot of confidence that models that are running on Cloudflare, that are running using our serverless platform, we’re going to be able to deliver the best price points for our customers while still having those be relatively low capex investments and relatively high margin for us. And again, that’s kind of Cloudflare’s bread and butter over the last 15 years. And finally, I think that inference and agents are the really big opportunities in AI. And while a lot of the money that’s been spent — has been spent on training for the last little bit, I think as we go forward, a lot more of that money is going to be spent on inference and agents.

And again, Cloudflare Workers is — the killer application for Cloudflare Workers is inference. It is agents. And we’re seeing more and more companies that are running with exactly that. So, I think DeepSeek is across the board a positive sign for Cloudflare.

And I think our team is excited not to be the only ones out there shouting that there’s a lot of efficiency that we can wring out of these models.

Hamza Fodderwala — Analyst

Thank you. Look forward to seeing you next month.

Operator

Up next, we’ll hear from Matt Hedberg, RBC.

Matt Hedberg — Analyst

Great. Thanks, all. My congrats as well. A lot of good stuff, good content today.

I wanted to focus on the product side. CJ is well into his role. The product team is built out. Matthew, you just spent on talking about a ton of innovation around killer AI apps and inferencing.

And there seems like there’s a lot of really, really exciting developments there. If you could just step back and sort of reflect on what you’re most excited about from a product innovation perspective this year. Are there things that we should be watching for inflection points this year?

Matthew Prince — Co-Founder and Chief Executive Officer

Yeah. I think I was just — I was actually at our regional sales kickoff in London with CJ and Mark this morning. And I think the first thing I’ll say is that Cloudflare has always been great at engineering. For, I think, at least the last seven years, we’ve been a great product.

What I think I’m most excited about is that 2025 is the year where we’re going to prove that we can be world-class and go-to-market. And what I think is unique about a product and engineering leader like CJ is he is so customer-focused. He’s so centered around making sure that our sales team has the tools, have the support that they need in order to win. And you could just feel that coming through with the team in EMEA over the last two days that I was there with them.

And I think that’s the thing that I’m most excited about. I think beyond that, I think that AI over the last year has felt like it went from kind of an interesting science project. It’s something that’s turning into real use cases. And I think that that’s going to be the turning point here.

And so, I talked a second ago about how inference and agents we really think are the big opportunities. Some of the applications that we’re seeing getting built are just really fascinating. In finance, there’s a financial service company that’s built an agent to perform — on Cloudflare Workers to perform diligence on M&A targets. In medicine, we saw a start-up that’s building an AI agent to review clinical trial results.

There’s an upstart customer support company that built an agent to pull a comprehensive view of customers and then make sure that that’s in front of the human agent that’s actually supporting a customer in a really efficient way. Again, all built on Workers. In CRM, a customer built an agent to provide an in-depth information about prospects, their businesses, the org structure in order just to make salespeople significantly more efficient. And again, all on top of Workers.

In media, an existing customer built an agent to answer questions about legal documents. In marketing, a customer built an agent about researching the latest topics and creating SEO design pages to promote their products. At Cloudflare, we built an incident management agent that’s saving us somewhere between four and six hours every single day. So, I think that — I know there’s a lot of hype around AI.

I think over the — what I — and I wasn’t quite sure how quickly that would turn into real ROI to customers. I think in the last couple of quarters, we’re starting to see where people are using Workers AI to deliver real value, not just something which is a cool demo.

Matt Hedberg — Analyst

Great. Great examples, Matthew. And then maybe just a quick one for Thomas. You mentioned pool of funds in your prepared remarks.

Is there any way that you could help us think about what the impact or quantify the impact of pool of fund deals in Q4? And when you think about its impact in 2025, how should we kind of think about guardrails for what seems like an increasing component of revenue?

Thomas J. Seifert — Chief Financial Officer

Yep. The activity around pool of funds in the fourth quarter was comparable to what we saw in the prior quarter, around nine percentage points. We’re excited by this. In a subscription model like ours, revenue is a lagging indicator.

And the fourth quarter felt like one of these inflection points where things are lining up where sales head count translates into activity, translates into a pipeline, translates into ACV that will be ratably recognized as revenue over the course of the year. And we saw a lot of these early pointers, all pointing in the right direction, sales capacity, sales productivity, large deals with customers north of $1 million, and pool of funds deals. The guardrails we put in place are a mix of how we compensate and how we incentivize because the art in turning these pool of funds deals into accesses is you have to get customers to consume product. And I think we learned a lot over the last 12 years — 12 months in dealing with these constructs and are quite confident in our outlook, how these contracts will be recognized ratably in terms of revenue over the course of the year, especially in the second half of this year.

Matt Hedberg — Analyst

Thanks a lot. Best of luck, guys.

Operator

The next question is Adam Borg, Stifel.

Adam Borg — Analyst

Awesome. And thanks so much for taking the question. I just wanted to touch on the announcement this week on FedRAMP High. I was just really curious what this may mean from a network architecture perspective.

And speaking with investors, does this mean anything around fragmenting your unique and homogeneous network? So, long way of saying is, does FedRAMP High changed that? And then maybe if you could give an update on your government business more broadly, just both within the U.S. and internationally. Thanks so much.

Matthew Prince — Co-Founder and Chief Executive Officer

Sure, Adam. I’ll take a stab at the first bit and a little of the second bit and Thomas may have some more to add. I think first of all, the federal business is important to us, and we want to make sure that federal customers who need protection can get that. And that means making sure that we can comply with what the requirements are.

But one of the things that was important to our product and engineering team was that we did that in a way which was still inherently Cloudflare. So, we didn’t rush to do something that would maybe result in some quick near-term revenue but would cause us years of pain in terms of our network and our architecture over the long term. We also didn’t want to make it a piecemeal, where some Cloudflare products were available under FedRAMP High, but other Cloudflare products weren’t. We wanted the entire platform to be doing that.

And so, as a result, I think we’ve moved maybe a little bit more slowly and more cautiously, but it’s exactly in respect to your question, which is we don’t want to fragment the network. And so, we have designed a compliance strategy that allows us to use our existing network and be able to meet the FedRAMP High requirements. That has been an effort in both directions. I remember very clearly some conversations with federal government officials where there were requirements like all machines had to be in the U.S.

And I said, do you really want us to import the denial-of-service attack into the U.S. before we stop it as opposed to stopping it out at the edge of the network. And they said, obviously not, that doesn’t make any sense. And I think they’ve been willing to work with us and figure out what the right process is.

But I think we’ve built a very, very, very good plan that allows us to run the same network largely across the board without having to fragment it the way that some of the other hyperscalers have or build a specific gov cloud. And that is, I think, what will allow us to continue to have success in the federal space while maintaining the elegance of Cloudflare’s network architecture where literally any server anywhere on our network is capable of performing any of the functions that we need. So, I think that that’s powerful. In terms of the federal business, the list of federal agencies that we work with continues to grow.

We continue to do more and more with them in the United States. Outside the United States, that continues to be really powerful as well. I mentioned that I was just in London, I was meeting, as part of our sales kickoff, also with the CTO of NATO, who is a Cloudflare customer and really excited for what we can do for them. And again, our team is very proud of the fact that we’re helping protect so many of these critical infrastructures around the world.

Adam Borg — Analyst

Awesome. Thanks so much.

Operator

The next question today is Jonathan Ho, William Blair.

Jonathan Ho — William Blair — Analyst

Good afternoon. How do we think about the security performance this quarter? And were there any standout product categories that you’re seeing? Congrats again on the strong quarter results.

Matthew Prince — Co-Founder and Chief Executive Officer

I think the theme for the last couple of quarters has really been that more and more of our large customers are buying into the overall Cloudflare platform. And so, it’s less about buying one particular feature and more and more buying the broad suite of everything that Cloudflare does. When we say pool of funds, that’s what we’re really saying. It’s that a customer isn’t coming and saying, I just want a WAF, or I just want a modern VPN replacement, or I just want that.

They’re saying, we have bought into Cloudflare’s vision for what a network architecture should look like. We don’t know exactly what’s going to make sense in terms of us deploying that. But over time, we see a significant portion of our network relying on the Cloudflare products. And so, we’re willing to commit a pool of funds and then draw that down as it makes sense in terms of our own deployment in order to deliver that.

And so, I think that that broad platform and a real true platform, all built on one unified architecture is what is giving us the ability to do these large pool of funds deals. And I think the theme has been largely that it’s a platform that has been the killer feature that people are adopting. If you push me more — and again, I don’t — still not yet from the revenue perspective, but I continue to see so many amazing things happen in the developer space. The fact that we crossed 3 million developers using the Cloudflare platform is terrific.

The pace of that growth is just accelerating. And so, I think that’s the area where there continues to be sort of the most excitement. And again, it feeds into the overall platform as well.

Jonathan Ho — William Blair — Analyst

Excellent. Thank you.

Operator

Next, we’ll take a question from Fatima Boolani, Citi.

Fatima Boolani — Analyst

Good afternoon. Thank you for taking my questions. Matthew, I wanted to go back to your commentary and your very emboldened stance about the go-to-market organization just in terms of higher aggregate sales capacity and accelerating productivity momentum into this year. So, the question for you is, this being Year 2 with Mark at the helm of the whole organization and kind of doubling back to some of your comments in the prepared remarks, what changes might you be considering from a compensation or incentive perspective to maybe balance some of that really robust pool of funds momentum but also engender more of that mass market appeal for your growth, your act, so namely the developer services and some of the Zero Trust services areas that could be a real tailwind for you? Thank you.

Matthew Prince — Co-Founder and Chief Executive Officer

I’ll start, and then I think Thomas will have more of the nitty-gritty details that he can share. I think that — so we’re just starting. I don’t think we’ve quite come up on Mark’s one-year anniversary, but it’s coming up pretty soon. And I think the amount that he’s accomplished in the year has been really amazing to watch.

Building up his leadership team, really just increasing the caliber and quality of the overall AEs, the SEs that are on the team. And then I think, incredibly importantly, identifying some of the real stars that were already on our team and making sure that they get the recognition and rewards that they deserve. Part of that is definitely around thinking about things like territories, being much more disciplined about how territories are assigned. Again, I was just hanging out with our EMEA sales team, and all salespeople always grumble about territory assignments at some level.

But all of them said, “You know what, honestly, this is the right way to do it, and it’s the fair thing to do.” I think one of the things that was sometimes a challenge, the blessing and the curse to Cloudflare is everybody needs Cloudflare, which is obviously wonderful. But if you don’t do territories right, the challenges you can have, a salesperson actually do quite well by selling one product to a bunch of different customers as opposed to going very deep and very specific and targeted on customers. I think Mark has done the things in order to deliver that. And then as we’re shifting to more pool of funds deals, we’re shifting away from being paid by what you sell, what your ACV is.

And we’re having AEs that are being paid based on the actual revenue that’s recognized as products are consumed. And again, I think that that’s important, as Thomas alluded to already, about making sure that our team is staying with customers, building deep relationships with them, making sure that they’re getting the most out of every product that they buy from us. And again, these are not — nothing here — we’re not reinventing the wheel in any way. I think we’re in a place where we’ve got great engineering.

We’ve got great product. And now, we’re doing the table stakes thing to become really world-class in go-to-market as well.

Thomas J. Seifert — Chief Financial Officer

Yeah. There’s not much to add to Matthew’s answer. I think one of the reasons why this transformation has been so successful so far is that we avoided major disruptive moves. And what is in front of us for the remainder of the year are tweaking at the edges.

I think Mark and his team have done an extraordinarily good job getting everything ready as we start the year, new territories rolled out, productive AEs in place, quotas assigned, a good handle around the account structure, and avoiding really major disruptions as we move through the year. There will be slight adjustments here and there, but all the major things are in place already.

Fatima Boolani — Analyst

Thank you both.

Operator

We’ll go to Gabriela Borges, Goldman Sachs.

Gabriela Borges — Analyst

Hey, good afternoon. Thanks for taking the question. Matthew, I wanted to ask you the go-to-market question in a slightly different way. I think about how consistently your market share for several of your products is higher at AI-native companies versus the rest, how do you navigate that conversation at the enterprise level where perhaps parts of the organization maybe pushing leading edge, but the rest of the organization may not be? And to what extent do you still get pushback on the depth of ecosystem on the Workers platform? And how do you address that? Thank you.

Matthew Prince — Co-Founder and Chief Executive Officer

Yeah. I mean, the whole AI space is so new that in almost any company that you encounter, even the digital natives, there will be AI evangelists and AI skeptics and they are mixed together. I think to start out, it starts with Cloudflare having a very self-serve, easy to use, easy for a small team to provision and experiment with low capital investment, like you can literally sign up with a credit card and start doing inference tasks and build a demo as a small team. And then we have let that bottoms-up approach work incredibly well for getting the adoption that we have in those use cases.

I think what changes in 2025 is that we’re going to be much more targeted with going after specific targeted accounts where we think there is the right fit where someone is already a Cloudflare customer, they’re using us in ways that are interesting. They know us. They trust us. And so, some of you have been to like our last Investor Day, Aly Cabral, who’s on the PM team, overseeing a lot of Workers, actually switched over to the go-to-market team, and she’s running what we’re calling a speedboat, which is targeting a handful of customers to specifically go after them and build that kind of library of case studies that are there, making sure that regardless of what size of company, regardless of what industry they’re in, that we’ve got those case studies that we can then use to expand over time.

So, I think the very nature of how Cloudflare is provisioned makes it very easy for us to go bottoms up in a way that doesn’t always require a big go-to-market effort, but then I think we’re being very strategic with our go-to-market team for the first time really in saying, let’s pick out the few hundred customers that we want to win and prove those as use cases. And I’m incredibly excited for Aly and her team and what they’re going to be able to deliver because she’s the right person, knows the platform extremely well, and is excited to go show people what they can build using Workers AI.

Gabriela Borges — Analyst

Yeah, that makes sense. Thanks for the detail.

Operator

Up next is Trevor Walsh, Citizens JMP.

Trevor Walsh — Analyst

Great, thanks for taking my questions. Matthew, you’ve talked a lot about agents within the context of AI today, both in your remarks and kind of some of your responses to questions. Could you just expand maybe just giving a couple kind of what in your mind are the two differentiated pieces or kind of product aspects of Workers that makes it well positioned for agents or why it’s, I guess, easy to build those there on that part of the platform? And then you spoke additionally about great examples of more start-up-type companies building agents to presumably go sell. But what about your larger customers kind of building agents internally, maybe for their own use? And is that something you’re doing on Workers? Do you have any examples of kind of that type of activity happening? Thanks.

Matthew Prince — Co-Founder and Chief Executive Officer

Yeah. I mean, some of the examples I gave you are actually pretty large companies. I think the difference between start-ups and large companies often is how much they let us talk about what they’re doing. But we’re seeing across the board really exciting things regardless of the size of the company.

Start-ups are sort of often just a little bit more willing to let us toot their horns. In terms of — I think the platform question is really important. And it’s important to think about like how to contrast Cloudflare Workers and Cloudflare Workers AI with what the hyperscalers are doing. So, let me first walk through.

Let’s say that you wanted to build an agent on a hyperscaler. So, the first thing is, you’ve got to contact them and say, “Listen, I need to use a machine, and a machine has to have some GPU resources.” And for various reasons that are pretty technical, you have to have a full VM basically to do that. So, you’ve essentially got to have a significant portion of the machine and in some cases, rent the entire machine. And in order to get the best pricing for that, kind of the competitive pricing, you have to commit to typically about a year of renting that machine.

So, you got a machine, it’s got an NVIDIA processor or more that are in it, you’re renting it for that entire year. It is then up to you to figure out how to maximize that utilization. And it was actually interesting, talking to CJ in some of the previous roles that he’s been in, and he told us what he thinks is common across most companies, which is that most of the time when you’re building anything with AI and inference, the level of utilization that you’re getting, it maxes out at about 10%. And what we hear from most people is that the average utilization is in the mid-single digits in terms of that utilization.

And it kind of makes sense because when these agents get invoked, you don’t know when people are going to submit queries or not, time of day makes a difference, all kinds of things make a difference in terms of when that availability is. And it’s up to you. The hyperscalers, their job, their business is renting you the machine, but after you have the machine, it’s up to you to get the most out of it. And so, that means that it can be very, very, very expensive to have the peak capacity that you need, but then waste all of that capacity when there’s downtime that’s out there.

And so, at Cloudflare, we’ve just always thought of it differently. We have thought that you should only pay for when you’re actually getting real work done. And so, we just — our business is not renting you a machine. Our business is selling you the actual inference task, When the agent is actually invoked and it’s doing work, you pay for that.

If it’s not, you don’t pay for it at all. And so, that allows you to actually have much easier start-up costs because you don’t have to go and negotiate a year-long deal with Microsoft or Google or Amazon, or Oracle. Instead, you can just come to us, put your credit card in, and only pay for what you use. We tend to have pricing which is still attractive.

And the question is then how because, again, we don’t get a significant discount on the underlying GPUs. But what we are able to do is drive up that utilization. And so, where a typical customer is seeing sub-10% utilization across their GPUs, our peak utilization of GPUs is now around 70%. The troughs are still much lower than that.

We’re much better still at CPU utilization, at getting an average up in the 70% range. But we see a path for us to be able to do that with GPUs as well. And what that effectively means is, again, roughly speaking, we can get effectively seven times today the amount of work out of $1 of capex spent in this area than you can with the various hyperscalers. And again, that either means that we can capture that as margin, or we can pass that on to our customers.

And so, at some level, what Cloudflare has always been is a giant scheduler. We are really good at moving data at very, very, very low cost, in most cases, even free around the world to wherever they are, the resources to be able to process that data. And that then in the agent space, in the inference space, in all of Workers, that is allowing us to have this incredibly high utilization and deliver just a much easier to get started with, easier to consume. You don’t have to have a whole SRE team to run your machines because we handle that all for you.

So, it’s easier to use. It’s more cost-effective. It’s fairer in terms of the pricing. It scales all the way down to zero when there’s nothing using it.

Scale’s up as much as you need on the other end. And we think it’s just a better model. What’s also interesting is because we’re inventing a lot of this IP and a lot of this tech ourselves, how you take a GPU, which was never intended to be a multi-tenant thing, and make it multi-tenant, we think that at least in the short to medium term, that’s going to give us a moat because there’s a lot of hard engineering work in order to deliver that. And we just don’t see any other companies that are focused on the same things that we are today.

Trevor Walsh — Analyst

Awesome. Super. Interesting. Appreciate it.

Operator

Our final question today comes from Andrew Nowinski, Wells Fargo.

Andrew Nowinski — Analyst

Great. Thank you for taking the question and squeezing me in here. Congrats on very strong execution again. So, I just want to maybe switch topics to your SASE solutions.

I mean, there’s no doubt that the network security market is transitioning to Zero Trust, but there are a number of vendors in the space. I’d imagine a lot of the bake-offs that you’re in are very competitive. So, I’m just trying to parse out maybe how much of your SASE wins are coming from the overall market inflecting higher with greenfield opportunities versus competitive displacements. And maybe if you can give some color on why you’re winning there? Thanks.

Matthew Prince — Co-Founder and Chief Executive Officer

I think when we first got into the space, we thought that the right strategy was to go after greenfield opportunities. But what we realized pretty quickly was a lot of customers that were out there were really dissatisfied with their — the sort of first-generation Zero Trust vendors. They didn’t like their performance. They tended to be — have low reliability.

There was a lot of arrogance in the market from certain vendors. And there was, I think, a feeling that there wasn’t actually kind of continuous innovation in the space. And so, somewhat to our surprise, we actually pivoted to going directly after a lot of people who are having problems with their first Zero Trust vendors that were out there. And I think where our pitch and where we end up winning time and time and time again is by saying we have the fastest network everywhere in the world, first.

Second, we have a comprehensive solution for your entire network, not just the Zero Trust piece, but for both the sort of forward proxy and reverse proxy parts and so we can bundle that all together. And then as a result, that bundle has a just much higher ROI than anyone else is able to deliver. And so that’s been a very compelling prospect. A lot of the big Zero Trust wins that we’re winning are displacement of other Zero Trust vendors.

We also are winning some of the new greenfield opportunities as well. And we continue to invest in the space. But the key here is because we have the broad platform, we’re able to just solve a much — a more diverse set of problems for a customer than just Zero Trust. And I think that over time, Zero Trust is a feature.

What people really want is a completely secure network, and Cloudflare is by far the leader in providing that soup to nuts.

Andrew Nowinski — Analyst

Perfect. Thank you, Matthew.

Operator

At this time, there are no further questions. I’d like to hand the conference back over to Mr. Matthew Prince for any additional or closing remarks.

Matthew Prince — Co-Founder and Chief Executive Officer

Thank you so much. I appreciate everybody tuning in. I wanted to take a second to thank our entire team for what was an incredible 2024. I think as Thomas alluded to, we see the first half of 2025 being a real turning point, and I am excited for what is ahead.

Thanks to the entire team. I’m excited to see a bunch of you over the weeks ahead as I travel to various sales kickoffs around the world and appreciate it and go Cloudflare.

Operator

(Operator signoff)

Duration: 0 minutes