Q2 2025 Earnings Call Transcript

Hello, and welcome to the Cloudflare Second Quarter 2025 Earnings Call. [Operator Instructions] I would now like to turn the conference over to Phil Winslow. You may begin.

Thank you for joining us today to discuss Cloudflare's financial results for the second quarter of 2025. 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, our 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 moments ago.

We would also like to inform you that we will be participating in KeyBanc's Technology Leadership Forum on August 12, Stifel's Tech Executive Summit on August 26 and Goldman Sachs's Communacopia and Technology Conference on September 9. Now I'd like to turn the call over to Matthew.

Thank you, Phil. We had an excellent quarter. We crossed $2 billion in annual run rate revenue, achieving $512.3 million of revenue in the quarter. We started the year detailing our strategy to drive reaccelerating growth.

Our Q2 results highlight that this formula is working and mark a key inflection point for the company, with revenue growing 28% year-over-year, up from 26.5% in the first quarter. We now have 3,712 customers paying us more than $100,000 per year, a 22% increase year-over-year. Revenue contribution from these large customers grew at 35% year-over-year, contributing to 71% of revenue during the quarter, up from 67% in the second quarter last year. Our dollar-based net retention was 114%, up 3% quarter-over-quarter.

Our gross margin was 76.3%, in line with our long-term target range of 75% to 77% -- we delivered an operating profit of $72.3 million, representing an operating margin of 14.1%, and we generated strong free cash flow of $33.3 million during the quarter, again exceeding expectations. Cloudflare keeps innovating faster than ever, and customers are voting with their wallets. You can see that in the momentum from our Q2 results. It's not just interest, it's real investments that drove record ACV bookings in the quarter.

Beyond innovation, under the leadership of Mark Anderson, our President of Revenue, we also delivered significant operational and strategic progress along multiple go-to-market areas in the second quarter. This sets a strong foundation for the rest of the year and beyond. Some highlights of this. First, the number of ramped account executives increased year-over-year at the fastest pace in the last two years.

We expect growth in our net sales capacity to continue to accelerate in the second half. Second, we delivered another year-over- year and quarter-over-quarter improvement in sales productivity. Third, we again saw particular strength with our largest customers, those that spend over $1 million and $5 million with Cloudflare annually, with both cohorts growing year-over-year at their highest levels since 2022. And finally, new pipeline attainment exceeded our expectation and grew at the fastest rate in more than two years. I once again feel like the company is firing on all cylinders.

The momentum you see in these results shows that we have the right technology, the right strategy and importantly, the right team to accelerate Cloudflare's next phase of growth. That's a good segue to discuss some of our wins in the quarter. A rapidly growing AI company expanded their relationship with Cloudflare, signing a 1-year $15 million pool of funds contract for Workers AI. This is the third contract signed with this customer in the last year as they moved all of their inference workloads from a hyperscaler over to make Cloudflare their single inference cloud platform.

The continued expansion with this customer demonstrates not only the tremendous value they realized from the Cloudflare platform, but also the truly unmatched scalability, efficiency and speed of Workers AI. Cloudflare is increasingly the platform the most innovative companies are choosing to power the future of AI. A Fortune 500 financial services company expanded their relationship with Cloudflare, signing two 3-year contracts totaling $11.4 million for application services and Magic Transit. This customer initially approached us looking to bolster their network resiliency with a dual vendor strategy, and we were happy to come in as the #2 behind their incumbent provider.

Impressed with our superior reliability, best-in-class performance and innovative products in just one month, this customer signed a second contract, making Cloudflare their primary vendor. A Fortune 500 multinational financial services company expanded their relationship with Cloudflare, signing a 3-year $7.1 million contract for application services, Magic Transit and Workers. This customer turned to Cloudflare to establish greater network resiliency by eliminating any single point of failure, migrating half of their traffic from a long-time incumbent to us. A Fortune 100 global financial services company expanded their relationship with Cloudflare, signing a 1-year $5 million pool of funds contract with initial use cases for Magic Transit, e-mail security, threat intelligence and application services.

In addition to addressing pressing reliability and redundancy requirements in order to improve their network resiliency, this customer was also able to enhance their security posture and gain unparalleled threat intelligence collected from our vast global network. A large state government entity in the United States expanded their relationship with Cloudflare, signing a 5-year $5.1 million contract for our SASE products, including Secure Web Gateway, Magic WAN, DLP and CASB. This customer's previous architecture was a mess of multiple vendors, including a first-generation Zero Trust vendor that was only 30% deployed after three years.

Consolidating on to Cloudflare's unified platform will improve the customer's overall security posture and simplify their architecture while also realizing a roughly 60% cost savings. A Fortune 500 technology company expanded their relationship with Cloudflare, signing a 3-year $2.4 million Zero Trust contract. What's neat about this deal is we initially lost this RFP 1.5 years ago to a first-generation Zero Trust vendor who was ultimately unable to meet the company's requirements, leading the customer to come back to Cloudflare. They were blown away by how quickly our Zero Trust products had matured in just 18 months.

This pace of innovation, combined with our ease of deployment and superior performance were key differentiators to securing this win this time around. A rapidly growing AI company signed a 5-year $4.6 million contract for AI Gateway, Magic Firewall, Magic Transit and application services. As a highly technical company, this customer turned to Cloudflare as a strategic partner to enable accelerated innovation, provide enhanced security, improve performance and offer unmatched scale with our globally distributed connectivity cloud. This contract is just the beginning with this customer.

They're already kicking the tires on our firewall for AI product. A leading digital travel company expanded their relationship with Cloudflare, signing a 4-year $3.8 million contract primarily for our Workers developer platform. This customer is transitioning workloads from an incumbent hyperscaler to Cloudflare Workers to drive faster innovation and better empower developers while also decreasing latency and improving their global end user experience. In the words of this customer, "The performance improvement we saw with Cloudflare was crazy".

This customer is a great example of our land-and-expand model across our products. They started with DNS in 2023, added application security and performance in 2024 and now are building on top our Workers development platform. What's next? They're currently testing our Zero Trust solution.

Some of our most strategic customer wins in the quarter, however, weren't big ACV deals. Let me explain. Cloudflare has historically had relatively low penetration in media companies. They didn't spend a lot or have significant security concerns, so they weren't our top priority.

However, over the last year, we've gotten to know their senior leaders at many of the leading publishers to understand new threats to their business. Historically, publishers online have made money primarily in two ways: subscriptions or ads. In either case, the key was generating traffic. In the past, one of the most effective ways to do that was through search.

Over the last 25 years, publishers allowed Google and other search engines to copy their content in exchange for sending them traffic. But recently, that traffic has been falling dramatically. Based on the data that Cloudflare has observed, it's nearly 10x harder to get traffic from Google than it was just 10 years ago. What's changed?

The interface of the web is switching from search to AI. Even at Google, which has represented the dominant interface for discovering the web, most searches now include an AI overview, which Pew Research has found significantly decreases the likelihood of someone clicking on a link and reading original content. Pew's data aligns exactly with what we've observed based on our customers' traffic. It's even worse with pure AI companies.

Every AI company we've tracked is worse than the Google of old with some being as much as 30,000x harder to get traffic from. As the interface of the web switches from search to AI, it's clear more people will read derivatives of content rather than the original content itself. That means the new AI-driven web will kill the old Webs business model. Cloudflare is in a unique position to help.

More than 20% of the web sits behind us today. But maybe as importantly, around 80% of the leading AI companies know and use us. So in Q2, we partnered with the who's who of the publishing world from the Associated Press to Ziff Davis and nearly everyone else in between to help invent the new business model for content creators on an AI-driven web. The deals we are signing with these companies aren't high dollars, but they are highly strategic.

The response has been incredibly positive from publishers for sure, but also from the majority of AI companies who understand that original content is the fuel that powers their engines. When seismic shifts happen in ecosystems as important as the web, new business models inherently emerge. We believe we are uniquely positioned to power the business model of content creation in the coming AI-driven web, but the opportunity may actually be much larger than that. The same rails that we are building to power payments from AI companies to publishers, we believe will be used to facilitate transactions between AI agents, whatever they happen to be doing for you online.

The fact that we sit in front of so much of the web and that more than half of our dynamic traffic is already between APIs means that we are strategically positioned to deliver the agentic web of the future. For those of you who have been following us for a while, you know that we talk about our product areas in terms of acts. Act 1 are our reverse proxy products, WAF, DDoS mitigation, et cetera. Act 2 are our forward proxy products, Zero Trust, VPN, network firewall.

Act 3 are our Workers developer tools. What we are doing to help publishers empower agentic transactions is a big enough deal to us that we've begun to refer to it internally as Act 4. Now you may not know this, but I was an English literature major in college with a computer science miner. I read a lot of Shakespeare, and all of his plays had five acts. So don't think we're done here.

We've still got a lot more up our sleeves. With that, I'll turn it over to Thomas, our CFO, who thankfully study economics, not English literature. Thomas, take it away.

Thank you, Matthew, and thank you to everyone for joining us. At the beginning of the year and again during our Investor Day, we detailed the factors that gave us confidence to drive reaccelerating growth over the course of 2025. We are pleased to have delivered on that goal during the second quarter with revenue increasing 28% year-over-year. As Matthew mentioned, strength in our business this quarter was driven by large $1 million and $5 million-plus customers continuing our momentum in the enterprise segment, green shoots across the financial services, public sector, retail and media verticals, continued momentum with our Workers Developer platform, including Workers AI and ongoing prioritization of security and resiliency by our customers.

In addition to accelerating the net capacity of our sales force, we also delivered another year-over-year increase in sales productivity, improved deal close rates and exceeded our expectations for new pipeline attainment. Turning to revenue. Total revenue for the second quarter increased 28% year-over-year to $512.3 million. From a geographic perspective, the U.S. represented 49% of revenue and increased 22% year-over-year.

EMEA represented 28% of revenue and increased 29% year-over-year. APAC represented 15% of revenue and increased 44% year-over-year. Turning to our customer metrics. In the second quarter, we had approximately 266,000 paying customers, representing an addition of over 15,000 paying customers sequentially and an increase of 27% year-over-year.

We ended the quarter with more than 3,700 large customers, representing an increase of 22% year-over-year. Revenue contribution from large customers increased to 71% of revenue during the quarter, up from 67% in the second quarter last year. We again saw particular strength in our largest customer cohorts, adding a record number of customers year-over-year spending both over $1 million and over $5 million with Cloudflare, which served as a tailwind to our expansion business. As a result, our dollar-based net retention rate accelerated to 114% during the quarter, up 3% sequentially and 2% year-over-year.

Moving to gross margin. Second quarter gross margin was 76.3%, representing a decrease of 80 basis points sequentially and a decrease of 270 basis points year-over-year. Recall that the extension of the estimated useful life of our network equipment from four to five years at the beginning of fiscal 2024, reduced depreciation for assets in service as of December 31, 2023, by about $5.6 million or 1.4% of revenue for the second quarter of 2024. During the second quarter of 2025, paid versus free customer traffic again increased as compared with both the year ago quarter and the first quarter, resulting in a higher allocation of expenses to cost of goods sold from sales and marketing.

At Cloudflare, we've always been clear our significant cost advantage is a strategic weapon. The accelerating adoption of our Workers development platform is a clear validation of this philosophy, demonstrating how the inherent scalability and efficiency of our network fuels our powerful engine of disruption. Even as we pass on substantial savings to Workers' customers compared with hyperscale competitors, we expect gross margin to comfortably remain with our long-term target range of 75% to 77%. Network CapEx represented 11% of revenue in the second quarter.

We continue to expect network CapEx to be 12% to 13% of revenue for full year 2025. Turning to operating expenses. Second quarter operating expenses as a percentage of revenue decreased by 3% year-over-year to 62%. Our total number of employees increased 18% year-over-year, bringing our total headcount to more than 4,600 at the end of the quarter. Sales and marketing expenses were $182.1 million for the quarter.

Sales and marketing as a percentage of revenue decreased to 36% from 37% in the same quarter last year. Research and development expenses were $83.6 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 $52.6 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 $72.3 million, an increase of 27% year-over-year compared to $57 million in the same period last year. Second quarter operating margin was 14.1%, a decrease of 10 basis points year-over-year. Operational excellence is a long-term competitive advantage, and these results highlight our continued focus on becoming more efficient and more productive.

Turning to net income and the balance sheet. Our net income in the quarter was $75.1 million or diluted net income per share of $0.21. Free cash flow was $33.3 million in the quarter or 6% of revenue compared to $38.3 million or 10% of revenue in the same period last year. We are comfortable with consensus free cash flow estimates for the second half of fiscal 2025. During the second quarter, we issued $2 billion of 0% convertible senior notes due June 2030. In connection with the offering, we also entered into a capped call option transaction with a cap price of 175% over the last reported sale price on June 12, 2025, which protects against dilution to a price of $469.73 per share.

We ended the second quarter with $4 billion in cash, cash equivalents and available-for-sale securities. Remaining performance obligations, or RPO, came in at $1.977 billion, representing an increase of 6% sequentially and 39% year- over-year. Current RPO was 66% of total RPO, increasing 33% year-over-year versus 29% in the first quarter and 30% for the fourth quarter. Moving to guidance for the third quarter and full year 2025. Entering 2025, data gave us confidence to invest to reaccelerate growth.

Second quarter results underscore that our strategy to deliver continued innovation and accelerating growth while also remaining committed to the strong unit economics of our business is working, and we are confident in our ability to continue to execute against this winning formula as we transition to the second half of the year and beyond. For the third quarter, we expect revenue in the range of $543.5 million to $544.5 million, representing an increase of 26% to 27% year-over-year. We expect operating income in the range of $75 million to $76 million. We expect an effective tax rate of 20%.

We expect diluted net income per share of $0.23, assuming approximately 376.5 million shares outstanding. For the full year 2025, we expect revenue in the range of $2.113.5 billion to $2.115.5 billion, representing an increase of 27% year- over-year. We expect operating income for the full year in the range of $284 million to $286 million. We expect an effective tax rate of 20%.

We expect diluted net income per share over that period to be in the range of $0.85 to $0.86, assuming approximately 370 million shares outstanding. In closing, we continue to focus on creating significant shareholder value with our ongoing commitment to disciplined execution, durable growth and operational efficiency. I'd like to thank our employees for their dedication to our mission as well as our customers for trusting us to help modernize, accelerate and secure their businesses. And with that, I'd like to open it up for questions.

Operator, please poll for questions.

[Operator Instructions] Your first question comes from Keith Weiss of Morgan Stanley.

Congratulations on a really solid quarter. It definitely looks like the engine is back to running full speed here. I wanted to dig into like the business model for the Agentic Web. And maybe, Matthew, you could give us a little bit more color and visibility on what that means in reality.

What are the business models that you're looking to enable for your customers? And how do you monetize that for Cloudflare?

Sure. Thanks, Keith. I don't think we know exactly the answer to that. And my hunch is that there will be a number of different models that emerge and over time, consolidate.

The analogy I've been thinking about is risk of hubris. When Apple rolled out $0.99 a song, that was a key turning point in the music industry, but it wasn't the ultimate model that we ended up with. We came closer to something that was $10 a month with Spotify. And so I think that this is going to go through a number of different stages and iterations.

And you could imagine something that is a fraction of $0.01 per transaction. You could imagine different sites charging different things. You could imagine sites that charge agents more or sites that actually discount for agents that are there. I think what we feel confident, though, is that because of the fact that so much of the Internet sits behind us and inherently, those agents are going to be passing through us that we have an opportunity to help define what those rails are that the agents will ride on and take some fee from that -- those transactions as we've helped facilitate them and make them faster, more reliable, more secure, give people the access to those rails.

So I think it's too early for us to model exactly what that looks like in terms of revenue, way too early. Right now, what we are playing for is very much around how do we just get as much adoption as possible? How do we make sure that we are the universal translator regardless of what protocol someone uses, whether it's MCP or the protocol coming out of Google or what's coming out of Microsoft. We want to make sure that no matter what it is that we work with it.

And again, I think that we're in a unique strategic position because of how much of the Internet does sit behind us.

Got it. Great to see the fast pace of innovation. And so we'll stay tuned on how these models evolve.

The next question comes from Andy Nowinski with Wells Fargo.

I extend my congrats as well on a great quarter. So we saw in the news how Cloudflare blocked a number of record-breaking DDoS attacks this quarter. And while your WAF, your DNS and your DDoS solutions are your Act 1 products, they seem to be seeing an inflection just like your newer Act 2 and Act 3 and Act 4 solutions. So Matthew, I also saw on X, that chart you posted about one of those massive attacks that consumed only a few percentage points of your network capacity while it consumed about half your -- the capacity of your competitors.

So I'm just wondering if you could maybe talk about the Act 1 segment of your portfolio and what's happening there?

Yes. We love the Act 1 products. And they, I think, are probably the easiest way to see the fundamental architectural advantage that Cloudflare has over really everybody else in the space. The way that most of our competitors try and deal with these problems is they set up specific scrubbing centers that have a certain amount of capacity.

Those scrubbing centers are not always optimized for the best performance. So traffic is not routed through them all the time. Only when an attack takes place, does it switch over. That means that, one, it inherently has a cost to route that traffic through.

Two, there's an inherent latency when you switch the traffic over because something has to change. And then three, those scrubbing centers just from a pure capacity planning basis, they have to be a limited size, and they have to keep up with whatever the sort of latest new attacks are. We took a very different approach from the beginning. And the way we've always talked about this is every single server that makes up Cloudflare's network is capable of running every single service.

And that's a really big deal that I think sometimes people don't appreciate because it's a fundamentally different architecture than anyone else in the market has in place. It took a lot more work, a lot more engineering to make that work. What that then means is that across all of Cloudflare's network, there are no scrubbing centers. Every machine is capable of dealing with WAF requests is capable of dealing with DDoS requests.

And what that means is that under normal circumstances, when we're not under a massive attack, there's a lot more traffic that's flowing out of our network because we're a cashing proxy than it's flowing in. And the way you pay for bandwidth is on the greater of in versus out. And so unlike anyone else, when we receive these attacks, not only do you have the capacity to deal with them, but they don't actually change the underlying cost structure of our business because even with these major attacks, it doesn't actually drive up our bandwidth usage. And so that fundamental architectural change where we have built the hard systems, the hard technology systems to be able to deal with any of our services being launched anywhere.

I think that shows up in terms of our ability to win customers in those Act 1 products. But that same architecture, that same work that we did to be able to stop those big attacks is also what allows when somebody is signing up for our Zero Trust services to make sure that they don't just have good service in major metropolitan areas. But if their CEO is on vacation in Rwanda, we've got facilities in Kigali, and we can deliver the Zero Trust services from there. If you deploy something with Cloudflare Workers, it will scale up because your code as a customer of ours also has the ability to run on literally every server across all of Cloudflare's network, and that lets us scale up and then also scale down incredibly quickly.

And so that architectural change is what has allowed us to win in that Act 1 product -- set of products. But it is that same architectural change that also is allowing us to win in Act 2 and Act 3 as well.

That makes sense. That's very helpful. Maybe a quick follow-up for Thomas. I think you said you surpassed $2 billion in ARR this quarter, which looks like you're still on track to reach that $5 billion target in FY '28. I'm just wondering if you could talk about the path you're on relative to your expectations.

I would say we are tracking well to our expectations. We were optimistic entering the year when we gave guidance for the year and reiterated during our Investor Day that the signs we are seeing in terms of the progress we are making with its success with large customers, pool of funds deals, variable revenue and especially the progress the go-to-market team is making in terms of increasing sales productivity and increasing sales capacity overall makes us confident that we would reaccelerate and now the second quarter is a good proof point to it. A large part of the performance in the second quarter was coming from pool of funds deals and variable revenue.

We are tracking well to our expectations there. So I would just say that we're tracking well to our plans and are quite confident that the momentum we have generated so far is going to play well into the second half of this year.

The next question comes from Matt Hedberg of RBC Capital Markets.

I'll offer my congrats as well. Matthew, in your prepared remarks, it was a really good update on some of the go-to-market improvements. It feels like that and it is some of the technology improvements are driving this reacceleration. I guess maybe digging in a little bit more specifically there.

How do you think some of these changes are impacting your ability to land larger deals? You gave a number of examples this quarter, and obviously, there was a large deal last quarter. And then maybe as a quick follow-up, could you give us an update on some of the partner momentum? And perhaps is that helping with some of these large deals as well?

Yes. Thanks, Matt. I think that we, for a long time, we're very much a product-led growth company, where we let the products sort of stand for themselves, and that was the primary thing that we did. And I think as a product-led growth company, you can get to big-ish deals, million-dollar deals, but it's really hard if you're signing $5 million, $10 million, $20 million, last quarter, we had our first $100 million deal.

It's hard to do that just with pure product-led growth. And so what I think I have been just really, really impressed by the work that Michelle, that Mark, that our entire go-to-market team has done to really build the relationships with buyers where they understand the total capability of the platform. And so every time you hear that somebody is signing up for a pool of funds deal, what they're really betting on is they're betting on Cloudflare, they're bidding on the broad product suite that we have, and they're betting on the ability for us as a team to continue to execute. And that, I think, is coming through quarter after quarter after quarter.

And it's been -- I think it's been just really astonishing to see how we upgrading our team now have the real go-to-market athletes to be able to go out, explain the value that Cloudflare has, explain how the ROI for our products are. And then they're leading that and then the great products are what follow behind. In terms of partners, Mark and the sales team have really sort of reoriented Cloudflare to be a partner-first sales strategy. And you can see that our partner -- growth from the partner sales channel is growing faster than the rest of the business.

I don't think that we'll ever get to 90% plus that you see from the likes of Cisco or even a Zscaler. But I do think that we were under what is the right level for us and that, that will continue to outpace the rest of innovation. And I'm spending a lot more of my time interfacing with partners understanding what their priorities are, making sure that we are a good partner to them. And that is -- just has to be key as we continue to go upmarket and sign those larger and larger deals.

The partners are behind many of the large deals that we have. They'll continue -- we will continue to prioritize them and the leadership that Michelle and Mark have brought in to run our partners organization is truly, truly, truly world-class.

The next question comes from Gabriela Borges with Goldman Sachs.

Matthew, I wanted to revisit your comment on pay per crawl, and specifically catalyzing adoption. So talk to us a little bit about what the friction points can be in some of these conversations, particularly conversations with the AI crawlers and the frontier models. And is the decision-maker for an Act 4 type product, the same decision-maker as an Act 1, Act 2 product? How do you build sponsorship across different parts of the organization to catalyze Act 4?

Yes. So I wasn't surprised that publishers were excited about what we were doing. And we literally haven't encountered a publisher that wasn't 100% all in on what we were proposing. And it's been amazing to build those relationships.

I was surprised by the reaction from the AI companies. I thought that they would kick and scream quite a bit more than they did. And quite the opposite. I think they all understand fundamentally that content, original content, valuable content is the fuel that runs their engines.

A way of thinking of this is there are really three legs of the stool that you need to have in order to be an AI company. You've got to have GPUs or TPUs or whatever that is. And someone like OpenAI reportedly spends -- it's like over $10 billion a year on that GPU access. You've got to have great talent that is -- understand this new area from a research and scientific perspective.

And we've all seen the headlines about how between Facebook and OpenAI and Apple and others that there's this real war for that talent and the AI companies are spending billions of dollars on salaries for that. But the third thing that you need is you need great content. You need great original content. And forever or not for quite some time, there's just been an assumption that, that will be free.

And in the world of search engines, maybe that was okay. But we aren't building search engines anymore. We're building answer engines. And the difference between a search engine and an answer engine is a search engine directs you to that content where you can go and the content creator can monetize it.

An answer engine answers without you having to leave. And so there has to be some value creation back to content creators that isn't just based on traffic. And again, with a notably few set of exceptions, the AI companies understand that. And I think that you can see that reflected in some of the comments that have come out of the major tech companies as they said, we have to make sure that we support the ecosystem.

The key point, though, and I think this is what is the most important work that we have to do. The key point is that there needs to be a level playing field. It can't be that one company has a unique advantage in getting content where others don't. And so what we are now really working on is making sure that as we figure out what the market looks like going forward for this, that it is a level playing field, that new start-ups have an opportunity to exist that just because you're a legacy provider doesn't give you some unique access to content that others don't have, that there's a way to make sure that if you're small, you pay less and if you're big, you pay more.

And in an ideal case, there are lots of sellers into this market, the content creators, and there are lots of buyers into this market. And if Cloudflare can help facilitate that, I think it's interesting. In terms of the other half of your question in terms of who the buyer here is, I think at first for pay per crawl, the answer is that the buyer is going to be either a relatively limited -- relatively limited set of the AI companies that are out there today or more likely, it's going to be actually sort of what is going on in terms of a transaction as a piece of content is accessed. And that's true whether it's access for training or whether it's access is part of delivering an answer as part of an answer engine that you see from some of the companies that are out there.

And so I think actually, it is a different sort of transaction, but it is one where we feel like we have relationships with the right people. We're having the right conversations. And again, my biggest surprise of the last several weeks has been that the AI companies actually are saying, yes, we need to figure out a way to support the ecosystem, but we need to make sure that there's a level and fair playing field. And again, I think that's a place where Cloudflare can help play a really pivotal role.

Yes, makes sense. The follow-up here is on media within the broader construct of publishing. And your comments that this has been underpenetrated in the past. Talk to us a little bit more about how media goes from being a less attractive vertical for you to a more attractive vertical, particularly given some of your peers have had challenges with renewals in the CDN space and pricing and negotiations and things like that.

I think it may be that this is so strategic that, that we really just won't optimize on how do we extract as much from media companies. I think we've looked -- if you look at Cloudflare's business writ large, not just media companies, one of the things that I think investors have often had questions about is why do we have a free service? Like why would we give service away for free? And there's a huge number of benefits that we get from that.

But it very well may turn out to be the case that the collection of free users using Cloudflare end up being more valuable than the collection of enterprise users using Cloudflare because that content, which is there is something which has unique access, that long tail of content with gems that are part of it is something that as AI companies need to build powerful systems that really represent the sum total of human knowledge, they need to have access to those gems. And if we can do something where you sign up for Cloudflare and it's less about you paying us and more about us actually helping you get paid, that's actually, I think, something which is incredibly powerful regardless of what revenue we're able to capture for our Act 1, Act 2 or Act 3 services.

The next question comes from Patrick Colville with Scotiabank.

This one is for Matthew. We had a very large foundation model vendor publicly call out Cloudflare as a third-party subprocessor. I thought was really interesting given the undoubtedly explosive growth that we're seeing in that category in 2025. So I guess not to go into specifics of that relationship, but can you talk about how Cloudflare can deepen its relationships with foundation model vendors? And then which products can Cloudflare sell into these foundation model vendors?

Yes. Thanks, Patrick, for -- I think what's an interesting question. As we've said before, our best estimate is that about 80% of the major AI companies are Cloudflare customers today. And they use us across a couple of different services, and I'll highlight two.

So the first is security. The challenge if you put up a foundational model is every time that somebody runs a request against that model, it has real cost to you and it's measured in not fractions of pennies, but often in pennies. And so if somebody who can find a way to run requests against your model at a very high volume or in a way that you can't control or in a way that is automated and not actually what your subscriber is doing or if they can find a way to do things like longer credit cards, the credits and the tokens on these AI models now act almost as a currency that allow people to take stolen credit cards and turn it into effectively cash.

All of those are unique security threats that make Cloudflare just a great partner for those AI companies that we can sit in front of. That, I think, is where most of them start with us. What we are finding, though, is that increasingly, because of the fact that we have deployed GPUs across our entire network and made it so that we can do inference as close as possible to their users as we are all going from seeing these ChatGPT-like systems as miracles and starting to take them for granted, there's a real need for them to get the best performance as possible. And one of the most effective ways of doing that is moving the inference closer to where the user is.

At the same time, increasingly, as we see regulations spring up around the world, targeting AI companies, they need to keep the inference tasks as close to users as possible to meet those regulatory needs. And so Cloudflare Workers AI gives them the ability to run inference tasks as close as possible to users. We would not be today the right place for one of the really massive LLMs to run because those, in many cases, will require multiple different machines working in coordination. It is a more complicated task.

But for smaller models, we're finding that Cloudflare is the best place for anyone who's building that to run that. And over time, we are investing in making our systems able to support larger and larger and larger models. And so I think that we are unique in being able to do inference on a global scale almost anywhere in the world. And that is a place where if the AI companies start coming to us for security, they quickly then learn that they can get benefits from some of our workers products as well.

Does that make sense?

Crystal clear. And congrats on that partnership. It's undoubtedly really exciting. I guess as my follow-up, can I just ask about one of the big news items of the week?

I mean clearly, earnings is a big news item, but the Palo buying CyberArk was the other one. Our thinking is that the strategic rationale is for power to have a play in Agentic AI security. I guess, can you remind us how Cloudflare thinks about Agentic AI security and whether that can turn the kind of, I guess, the financial meter at some point this year? Or is it more of a kind of 2026 thing?

Yes. I think, first of all, you can't dispute that Palo Alto Networks is just an iconic company and doing a great job. Mark Anderson built a lot of their go-to-market engine for us, and they're a company that we have long looked up to. They have a very different strategy for R&D than we do, where a lot more of their R&D is through acquisitions.

And what we hear from customers is that when you try to stitch together a bunch of things, you end up not really with a platform, but with a Frankenstein. And that creates actually scams and gaps in security. It also just makes -- it makes for a very sort of complicated and expensive set of systems to try and stitch together. For us, we really -- we'll never say never that we won't do a big acquisition.

But I think that we really have an incredibly high hurdle rate for anything that we do in terms of acquisitions. And we really believe in internal innovation and internal R&D first and foremost. And so everything -- when we talk about how we can power the Agentic web, security is inherently going to be a big piece of it. And those agents because so much of the Internet already sits behind Cloudflare, those agents are going to flow through us.

And so providing the guardrails, providing the rule enforcement, those are all products that we already have in market that are there and waiting for as these systems develop. And again, I think that the problem isn't the products. The problem is just -- we're all living a little bit in the future of what this is going to be. But we have the technology, we have the products, and we have got that through our own internal and seamless development as opposed to through a series of R&D acquisitions.

The next question comes from Adam Borg with Stifel.

Maybe just for Matthew on Act 2 with SASE. So it was great to see you moving into the Visionary Quadrant, Gartner single vendor SASE Magic Quadrant this year, and you talked earlier about a win-back deal that you lost 18 months ago. So maybe just big picture, talk a little bit more about Cloudflare One competitively, what you're hearing from customers and are seeing in the market.

Yes. I think that when we think about Gartner and Forrester, our strategy, like I tell the team regularly, if the very first time we appear on one of those charts were on the top right that we waited too long to launch the product. We like to get products in the market as quickly as possible, get user feedback from our broad set of users and innovate faster than anyone else. And so if you measure effectively the hypothesis of any of the change for us in any of these charts, you see that we have -- that's sort of the fastest way to measure Cloudflare's innovation.

And I'm proud of the team at how rapidly they're innovating in this space. We love the fact that when customers give us a chance in the space that we have a great ability to win those customers because we're faster, because we're easier to deploy, because we can deliver a much higher ROI to customers. When we're in the mix, we like the fact that we win. And oftentimes, we're winning back deals that we earlier had lost or we're taking deals actually away from those first-generation SASE or Zero Trust vendors that are in the market.

And so I think what has held us back has been really awareness and both the partner-first motion, which Mark and the go-to-market team are putting in place as well as things like showing up in Gartner and Forrester, where today, we're now neck and neck with Zscaler in the space. I think that's the indication of how we're going to perform. And it reminds me a lot of with our Act 1 products, when we first launched, people would say, oh, you'll never be able to catch Imperva. They invented WAF, you'll never be able to do that or you'll never be able to catch Arbor or whoever was in that space.

We're just really good at innovating. And I think over time, that plus an underlying cost advantage is the recipe for winning in a category.

That's great. And maybe just as a super quick follow-up. Obviously, given the 20% of the Internet you cover, you've talked in the past about the crystal ball that you see. Any commentary on the macro that you're seeing?

I think that I think that it's -- that there is somewhat of just a disjointedness out there where we're seeing places where there's really strength in performance, and we're seeing others that are struggling. And I'm -- I feel very fortunate that even as we have kind of a very uncertain world, people are continuing to see the value that Cloudflare can bring, see how we can accelerate their businesses, deliver real security. And I think that we continue to be able to perform regardless of what the crystal ball shows.

The next question comes from James Fish with Piper Sandler.

Nice quarter. But I know we talk about AI all the time, but maybe just shifting off of AI a little bit. Is there a way to think about more of the broader Workers platform directly in terms of how you guys are thinking about capturing workloads that had previously been on a hyperscaler or on-premise versus kind of brand-new workloads? What are you seeing on that Workers core side of things?

Yes. Jim, it's there are a certain set of workloads that just work really well on us. And what we found is oftentimes the best way to get our foot in the door is not to say, hey, move your whole application, but to say, take one function that's a part of your application that is particularly mission-critical or particularly latency sensitive and move that one function over. And so that then tends to get people familiar with the Workers platform.

And once they see how they can save significant amounts of money, they can get much better performance, they can have inherent security from it, even if it's just moving one function over out of a larger application, that lets us have a toe in the door that lets us build that relationship over time. And I think that what we've learned is that as we do that, it then becomes easier for us to say, okay, let's look at some of the really big spend that's out there and how we can transition that away from the hyperscalers. And so we think that with our sort of Act 1 products, you can get 7-figure deals with our Act 2 products, 8-figure deals.

With our Act 3 products, we really think that there are a lot of 9- figure deals that are out there as we are able to take from hyperscalers or as Thomas likes to say, the best way to win a $100 million deal is to have it save the customer $100 million off what they were spending with one of the hyperscalers. And that, I think, is going to be a big driver of us winning more of those workloads. As we talked about at Investor Day also, we've -- Mark has formed a speedboat with Aly Cabral, who previously ran Workers on the product side. She said that she really wanted to go and dive in and run this from the go-to-market side.

And she and the team that she's assembled are just extraordinary at targeting specific customers, helping them see how they can get benefits moving away from one of the hyperscalers to Cloudflare and winning more and more of those customers. And so I think that's something that's really been working extremely well over the last few quarters.

Got it. And just as a follow-up, I know you guys talked about the state win, but we're all starting to think about the end of the federal cycle here and the impact of those. So I guess what are you guys seeing on this FedRAMP aspect that you guys have had for a few years now? What are you guys seeing in terms of trends and what that could look like heading into year-end?

Yes. So FedRAMP is a priority for us. We're on track to get to the point that we've met all the requirements before the end of this year. And once we do that, we can -- we are unrestricted in terms of the business that we can go after in the federal government.

And the final question comes from the line of Roger Boyd with UBS.

I wanted to come back to pool of funds. And I'm just wondering if you could provide a little more color around the trends you're seeing both from a bookings perspective, but also consumption perspective. And -- just trying to level set where we are in terms of the offsetting headwinds and tailwinds. It feels like it's becoming more of a net tailwind here, but would love to get your perspective.

And I guess the second part of it is the standout metric to me was a 3-point improvement to dollar net retention in the quarter. To what degree should we attribute the success there to some of the ramping consumption against pool funds?

Yes. Let me get started. So pool of funds deals with our largest customers represented low double digit in the second quarter, up from less than 3% a year ago. So significant progress.

As we get more mature, we get better visibility in how we track. I would say across all pool of funds deals, we are tracking on, if not slightly ahead of target. And one of the results you see because of that is that variable revenue coming from consumption of existing customers was a good contributor to the outperformance in the second quarter. And with that, the tailwind we saw in D&R.

We've been talking about bottoming out. We have seen that, that would happen. And now with the superior execution on pool of funds and us getting more comfortable with this go-to-market instrument, we are seeing the -- we are harvesting the tailwinds of that business.

And Roger, I don't have anything to add other than I appreciate you asking a question that Thomas could answer, so I wasn't alone studding and fading up here on the stage.

This concludes the question-and-answer session. I'll turn the call to Matthew Prince for closing remarks.

I'm incredibly proud of the entire Cloudflare team and how we've been executing over this just really interesting time. Lots of things are changing in how the Internet works, lots of things are changing in terms of AI and the fact that we can be at the center of all of these amazing trends and help shape what that future looks like is something that just makes me more excited about Cloudflare's future than I've ever been. So thank you to the entire team. Thank you to our customers.

We'll see you all back here next quarter.

This concludes today's conference call. Thank you for joining. You may now disconnect.