CEOInterviews.AI
Start App
Get an email when Matthew Prince gives a new interview
One email per new interview. Unsubscribe in one click.
Matthew Prince
Co-Founder, Co-Chairman & CEO, Cloudflare

Cloudflare CEO: We're Ready To Block Millions of Websites From AI — With Matthew Prince

📅 Sep 02, 2026 Alex Kantrowitz 51 MIN 506 VIEWS 66 SEGMENTS · 2 SPEAKERS
Matthew Prince is the co-founder and CEO of Cloudflare. Prince joins Big Technology Podcast to discuss Cloudflare’s plan to block Google from millions of ad and subscription-supported websites unless publishers choose to opt out. Tune in to hear why he believes Google must pay for access to online content, how AI is destroying the web’s traditional traffic model, and why the era of SEO is coming to an end. We also cover micropayments for publishers, the risks AI agents pose to small businesses, and how commerce could consolidate around a handful of giant companies. Hit play for an urgent look...

What Matthew Prince said

Written from the verified transcript and checked against it. Every figure links to the moment it was said.

Matthew Prince, CEO of Cloudflare, said automated traffic surpassed human traffic on the web in May 2026, earlier than his prior estimates of late 2027 or early 2027. He projected automated traffic could be a thousand times human traffic within five years, driven by agents and AI interfaces. Prince argued the web is growing again since October 2025, with more non-technical people creating content using tools like Lovable and Replit, making the internet healthier, not worse. He discussed the need for new business models, including micro-payments for content, and said Cloudflare will block Google by default for ad-supported and subscription-supported sites starting mid-September, treating Google like other AI companies. He expressed optimism for media companies, predicting 2026 will be a year of real shift where content creators get paid. He also worried about small businesses in an agentic commerce world, fearing consolidation, and is working with partners like Shopify and PayPal to address it.

Key takeaways

  1. Automated traffic surpassed human traffic on the web in May 2026, and could be 1,000x human traffic in five years.
  2. Cloudflare will block Google by default for ad-supported and subscription-supported sites starting mid-September, potentially dropping millions of sites from Google's index.
  3. Google sees more than 2x the web as OpenAI, 4x as much as Anthropic, and 5x as much as Microsoft, giving it a crawling advantage.
  4. Micro-payments infrastructure would need to scale to over 100 million financial transactions per second to support agent commerce.

Numbers and commitments

FigureWhat it refers toTypeAt
May 2026 Month automated traffic passed human traffic timeline 1:44
40 times Additional CPU production needed for one agent per knowledge worker metric 16:29
100 million Financial transactions per second needed for micro-payments metric 23:18
2x How much worse publisher crawl-to-click ratios got in a year metric 23:18
2x Google's web crawling vs OpenAI metric 26:31
4x Google's web crawling vs Anthropic metric 26:31
5x Google's web crawling vs Microsoft metric 26:31
$3.17 Cost to host a WordPress site price 13:10

Chapters

  1. 0:00Automated traffic surpasses human traffic
  2. 3:48Web growth and content creation
  3. 9:47AI traffic composition and load
  4. 12:55Infrastructure challenges at 1,000x
  5. 19:35Publisher economics and crawl ratios
  6. 23:18Micro-payments and new business models
  7. 26:31Blocking Google by default
  8. 38:27Small business risks in agentic commerce
  9. 47:37Solutions for small businesses

Questions asked in this interview

12
  1. 0:00So, what is going to happen to the web?
  2. 5:50So, yes, there's been this ability to create a lot of new stuff, but it also seems like a lot of that—I mean, I know you said it's not slop, but something close to it? What do you think?
  3. 9:47What are the proportions, if you have any sense of that?
  4. 16:03What makes you believe that that's going to be the case?
  5. 18:42So, just to put a fine point on it, the increase is going to be agent-driven, right?
  6. 22:11Where do those numbers stand today?
  7. 28:35... allow them to block the AI crawling, they're now doing the sort of negotiating in public where they're telling Google, "Hey listen, if you don't give us the terms that we want, we are just going to use Cloudflare's tech and block you"?
  8. 30:38And so what does that do to, let's say, a publisher's ranking in the search algorithm?
  9. 36:27So, you think Google's going to pay up before this deadline?
  10. 36:46How many of your clients do you think will actually go ahead with the blocking of these AI crawlers?
  11. 38:27So, can you explain sort of what the stakes are for small businesses again as this trend continues to progress and we go to a web that's so much more bot than it is human?
  12. 48:26... that a small business has a portal in the back end of ChatGPT and the small business owner spends their weekends dictating all the things about their business, or what could it look like and what are the solutions you've discussed so far?
Host 0:00 ↗
In an internet where AI bots now outnumber humans, what happens to the web and can news information publishers and small businesses survive? That's coming up with Cloudflare CEO Matthew Prince right after this. Welcome to Big Technology Podcast, a show for cool-headed and nuanced conversation of the tech world and beyond. Well, on the web, AI bots now outnumber humans. So, we have more AI traffic than human traffic and that's obviously changing the dynamics and the incentives for news publishers, information publishers, and of course small businesses. So, what is going to happen to the web? Well, today we are lucky to be joined by Matthew Prince. He is the CEO of Cloudflare. He was with us almost exactly a year ago talking about Cloudflare's efforts to prevent the web from disappearing. The web is still here, but maybe not as healthy as it was a year ago and obviously not due to the failure of those efforts, but the force of AI marches on and today we will talk about what that battle is going to look like. Matthew Prince, Matthew, it's great to see you again. Welcome back to the show.
Matthew Prince 1:03 ↗
Thanks for having me on.
Host 1:05 ↗
All right, let's go with the headline here, which is that recently Cloudflare has found that AI bots outnumber humans browsing the web. I can't tell whether that's surprising or not. I think part of me would say, "Hey, you know, of course that's happened and it's surprising that it just happened." The sort of conventional wisdom has just always been a lot of bot traffic on the web, but just talk a little bit about the escalation in AI traffic that you're seeing, whether that's, you know, crawling for new information or agents taking action and what it says about the rise in in these type of AI applications recently.
Matthew Prince 1:44 ↗
Yeah, you know what? I'm certainly surprised. I was asked at Web Summit in November of 2025 when we thought that automated traffic would pass human traffic online. And we pulled all the data and looked at it, and our best estimate based on trends was that it was going to be in the second half of 2027. At South by Southwest in March of 2026, I was asked the same question again. We pulled the same data again and we're like, "Huh, that's interesting. It's moved up. It's now we think going to be in the first half of 2027." And so I was floored when my team came to me and said in May of 2026, "You won't believe it, but we now see that automated traffic has passed human traffic for the first time in the history of the internet." And I think that that really does just show how fast this is progressing and how fast actually the derivative is changing, where the slope of the line is getting steeper. And if we just assume the current trends, what blows my mind is that if you just fast-forward 5 years from now, we estimate now that automated traffic online will be a thousand times what human traffic is. And that's not because human traffic is going to drop, it's going to be because automated traffic is just going to continue to rise like crazy over that period of time. And that's going to put, you know, enormous strain on the resources of the internet. It's going to put enormous strain on the infrastructure. And it's a lot of what the team at Cloudflare is thinking about is how do we make sure that the internet is ready for what is going to be, you know, just a complete and total onslaught of automated traffic.
Host 3:25 ↗
Right, and I'm sure you've heard of dead internet theory, but for those who haven't, it's this theory that one day the web is basically going to consist of bots and automated content and the bots will be manipulating that content for algorithmic curation. So what you see in your social feed is effectively a dead internet that looks like it's real. This is obviously a little different than that.
Matthew Prince 3:48 ↗
Yeah, actually the trends we're seeing are kind of the opposite of that. It's actually interesting. Cloudflare started and launched in 2010, and around 2015, the web stopped growing. In fact, it plateaued and then it was even declining slightly over the last 10 years. And then right around kind of October of 2025, the web started growing again and it started growing at a rate which most closely resembled kind of the mid-2000s. And the content that's being created is not like—I would not describe it as automated sort of slop. There's some of that, but there's always been some of that. What I think has happened instead is that people who were not particularly technical, people who didn't have a lot of resources, but they had an interesting idea, they had a story to tell, they had something to share, all of a sudden were able to use incredible tools like Lovable or Replit or what the team over at Base44, Wix is doing, and then have those tools be able to build pretty sophisticated applications. And that's like, for the first time almost in Cloudflare's history, we're seeing an explosive growth in things which are getting created. And instead of it making the internet worse, I actually think the internet's getting better right now. It's just the interface that we're all going to consume it through is going to be very different. It's going to be much more mediated through some sort of an AI interface. But I think it's incredible that even if you don't have technical skills, you can now put together an application, put together a website and bring that online. And again, I think that's remarkable and it's actually making the web—the web is getting healthier, not less healthy right now.
Host 5:50 ↗
Okay, I actually want to press you on this one because one of the studies that was done about this showed that, for instance, while there are many more apps on the Apple App Store, for instance, the reviews are going down, right? Seems like the weekly active users hasn't really increased. So, yes, there's been this ability to create a lot of new stuff, but it also seems like a lot of that—I mean, I know you said it's not slop, but something close to it? What do you think?
Matthew Prince 6:20 ↗
What I see every day is just people who didn't have the resources or the ability able to create things which are valuable. And so yeah, is there going to be stuff that's not valuable? Of course. But I think that the net has been that we're getting more things out there. I do think that we're going to move more and more away from kind of an App Store model because that's not going to be the interface where we reconnect with it. And the App Store in some ways was an enormous threat to the web. It took people away from the web. It was probably one of the biggest forces, between it and Facebook, were like the two biggest forces that really made the web less interesting. I think now putting things back on the web, and then making sure that they're accessible across the different AI systems in a way that makes sense for creators and makes sure that creators get compensated is going to be something which I actually think will revitalize the web, and it's exactly what we've been seeing. And again, I have both just personal examples from colleagues, family who have been able to get online and sell things, but then also professional examples where every day I'm getting letters from people who are like, "I'm able to do something that I just never thought I would be capable of doing before," and using some of these again incredible tools that are out there combined with the infrastructure that someone like Cloudflare provides very cost-effectively. I think it's actually making it so that humans assisted with AI are more creative than they've ever been.
Host 7:55 ↗
Okay, I'm just going to pause to say when I built the website for our summit, which happened in June—and when I'm saying "I," I should really put quotes around it—it was mostly the Claude Code harness hooking into the Cloudflare agent that went out and set up all these accounts for me and then built the site. So—
Matthew Prince 8:15 ↗
I mean, it's extraordinary. Compare that with what you would have had to do in the past, which is hire a whole—
Host 8:23 ↗
I wouldn't be able to. There's no way I could even start.
Matthew Prince 8:24 ↗
Like, you alone wouldn't be able to start, so you'd hire someone to do it. And that has enabled you to do more interesting things. And I think that's what technology at its best does: it empowers more people to have access to something than they had before. We're doing this remote. Even just 10 years ago, what would it have cost to do like a live video setup with 4K video in two different places? It would have been hundreds of thousands of dollars just to have the studio time or set up cameras in order to do this. And yet we've been able to do this. I think AI is no different. It's going to enable more people to be able to be creative and do interesting things. And if we get the incentives right, then hopefully it also makes it so that we can have more entrants into the marketplace. Figuring out what's the business model—I mean over the next 5 years, figuring out what the business model of the next 30 years of the internet is—that's the most interesting and important question in technology today, by far.
Host 9:33 ↗
Right.
Matthew Prince 9:34 ↗
If we get that right, we're about to be on a golden age of content creation, and it could be incredible. If we get it wrong, it's easy to list what the parade of horribles could be.
Host 9:47 ↗
Yeah, and so let's dig into this, right? Because there's two sides of it: there's the content creation side and then there is the browsing of that content. And this is sort of picking up a little bit about what we spoke about last year. More and more that content is being crawled or accessed by AI, right? So you've seen this massive increase in AI traffic to websites in the past year to the point where it's now overtaken humans. Do you have any visibility into whether that is AI bots accessing it to train new models primarily, agents using those websites as a human would, or maybe a query at test time? So as the bot is trying to find an answer, it goes to a site and it ends up bringing that back into the chatbot interface without a person actually going to the website itself. What are the proportions, if you have any sense of that?
Matthew Prince 10:50 ↗
It's all of the above. And so I think all of those will drive things. There's definitely an uptick in training as more model providers are coming online. OpenAI is crawling a little bit more than they did a year ago; Anthropic is crawling a little bit more than they did a year ago. But it's more that we've just got more people building models, and I think that's great. The world shouldn't be five frontier model providers; it should be 500,000. And how do we make it so that that's available? At the same time, you've got sort of agents that are operating on their own. That's probably the smallest of the categories right now. I think that that will grow and become a pretty large category. But right now that sort of agent-to-agent kind of commerce and things is more of a future thing than a present thing. And then I think you are getting more of the AI acting almost on behalf of humans going out and accessing things. But it's almost wrong to say that they're acting exactly like humans do. The scale is just dramatically different. If I'm shopping for a digital camera unassisted by an agent, I might visit—if I'm really interested—five websites, and on every website maybe visit, I don't know, 10 pages or something like that. My agent is going to visit a thousand websites and visit every page on all of them. And so the load that that then puts on this, especially if there's no financial disincentive to just crawl everything, is pretty substantial. Now, I think you've got to balance that against what's the benefit to the user in the fact that agents are infinitely patient, that they will look through everything. I think that's really good for end users. But there's servers behind every one of those requests, there's bandwidth that has to get paid for, there's content that had to get created. And so there's got to be some exchange of value. And if the AI companies are just taking without returning some sort of value, then that just doesn't make sense. And so that's again what I think we've got to figure out.
Host 12:55 ↗
Right. So I think when you talk about this, your view is that we're eventually going to go to—I think you said this—a thousand times more bot traffic than human traffic in 5 years, which is crazy. Okay, first of all, just—
Matthew Prince 13:10 ↗
To give you a sense of how intense this is: the European internet came this close to completely collapsing in COVID. Because in COVID, over a period of 2 weeks, internet traffic doubled, and especially doubled for video streaming, which was a very high bandwidth. I remember being on a phone call with senior leaders at Netflix, Disney, Akamai, YouTube, all the major streaming platforms. And it was basically like this Mexican standoff where nobody wanted to reduce the quality of their video stream, because they're like, "People pay for us because we have 4K video, and we're not going to reduce it." But what was happening was in Europe, everyone was putting their kids in front of the TV streaming first thing in the morning, and that was unfortunately lining up with Asia where everyone after dinner was turning on something to watch, and it was just causing so much traffic that we were running out of capacity online. And eventually we had to cut their video streaming resolution down in order to make sure the European internet stayed online. And by the way, if the European internet had fallen over, then Asia would have been next; the US would have been okay for a little bit longer. But at some level, there's a finite capacity to all this stuff, so you've got to be able to support it. And that was at 2x. At 1,000x, it's going to be completely different. And so again, we're spending a bunch of time trying to figure out how do we make sure that the technologies are ready. WordPress, the most popular platform for hosting content, is run on a programming language called PHP, not the most efficient programming language. And so we're like, "Listen, we want to support the WordPress community. We want to make sure WordPress is successful." But I was meeting with one of the largest WordPress hosts a little bit ago, and they said, "It costs us about $3.17 to host a website on WordPress, and then we charge like $12.99," and that's sort of what their business is. I was like, "If traffic went up a thousand X, what would happen?" And they're like, "Well, then it would cost us over $3,000 to host that website." And again, the average WordPress user isn't going to be willing to spend over $3,000 to host their WordPress website. So, we've got to change some of these fundamental technologies. And again, I'm thinking the best case: it's still the user interface and the ecosystem and everything that people have come to love about WordPress, but with the underlying technologies made much, much, much more efficient so that you can actually scale to what's coming. And that volume of what's coming is just—again, it's almost inconceivable to think about.
Host 16:03 ↗
All right, we have to spend some time talking about this, because this is very interesting. First of all, to anticipate a 1,000x increase in web traffic coming from bots in 5 years—I mean, yes, you've seen quite an increase recently, but a thousand X is a crazy number. So, what is going to make up that thousand X? What makes you believe that that's going to be the case?
Matthew Prince 16:29 ↗
We just look at the trends right now. In May, we saw that automated traffic crossed it, and we kind of extrapolate out what that rate of change is and then we assume that the exponent is going to increase on that, because it's still way underpenetrated with what's there. But we've got to invent new technologies in order to support it. So, just to give you some substance: let's imagine that just every knowledge worker on Earth had an agent that was working for them 24/7 helping them do their job. And again, that seems reasonable. Every knowledge worker is going to have some agent that's working for them. Very few do today, really, but over time that becomes reasonable. Just with that, and putting GPUs aside, just looking at CPUs: if every one of those agents ran in a traditional container, which is not even a VM, not even its own server, but just a container, you would need more than 40 times the current CPU production in a calendar year in order to support that level. And that's with one agent. I would guess that most knowledge workers are going to have many agents that are going to be working on their behalf. Using the current technology, we don't have the ability to support what's coming. And so, again, what we're spending a lot of time at Cloudflare thinking about is, like, okay, we went from servers to VMs to containers in order to get more efficient, but even containers, you've got to pull in an entire operating system, you've got to pull up all the toolchain, all this stuff. It's pretty heavyweight. So, what's even lighter-weight in order to support what's coming? A lot of what we've built with our developer platform, which we call Workers, and leaning heavily on technologies like Isolates that allow you to have much, much, much lighter-weight services—we think that those are the sorts of things that the internet's going to have to have if it's going to be able to continue to scale to support the level of agent traffic that just is inherent to what I think the future holds.
Host 18:42 ↗
So, just to put a fine point on it, the increase is going to be agent-driven, right? That 1,000x increase is agent-driven.
Matthew Prince 18:49 ↗
Or humans interacting with their agents, right? But it's just going to be automated. And the way to think about it is: if you've used ChatGPT, if you've used Claude, if you've used Grok, and you play with them and you just ask them a question, just count how many websites—you ask it a simple question, and it just goes across everything. Now imagine that not in the hands of probably about 1% of global knowledge workers in terms of using it in a regular basis, just scale that up to 100% of knowledge workers. And instead of it being something that you use maybe once or twice a day, it's something you're using for every task. And it's really easy just to see how a thousand X in 5 years is not only conceivable, but it might actually be a low estimate.
Host 19:35 ↗
Yeah, that is crazy. So basically, last year we talked about the economics for publishers might not make sense because if you're paying somebody to write an article and Anthropic crawls it and drops it into Claude, then they sort of get cut out of the value proposition.
Matthew Prince 19:51 ↗
That's right.
Host 19:51 ↗
But now a year later, as we've seen this rise in agentic traffic, the problem actually is much broader because now websites that you might not have even needed to invest a lot into making the content for, if they get faced with this increase of traffic, could cost much, much more to upkeep. And then the economics of even sustaining them become questionable if you can't find a way to do it more efficiently.
Matthew Prince 20:17 ↗
The thing that's really important is that different content creators, different publishers, different app developers are going to have different preferences. And at Cloudflare, what we believe is we want to make it as efficient as possible for the companies that want their content in AI, and we want to create the systems and incentives for those companies that don't want their content in AI, especially if they're not compensated for it, to be able to control that. But let's talk about the majority. The majority of companies that are out there, the majority of content that's created, they actually want that content in AI. Like Cloudflare, our knowledge base is a way of us sharing how to use Cloudflare really well, and we want to make sure that every AI system has it and has access to it. And so for that, what we have to do is figure out how to make it as efficient as possible. How do we make sure that we're not delivering a big heavy HTML page, but just a markdown that is actually required? How do we make sure that we can actually get systems into AI as efficiently as possible with caching and other things that are out there? Our goal has to be: how do we make every time we serve a page 1/1,000th of the cost of what it used to be? Because that's the only way that we're going to stay ahead of the demand which is coming down for those companies that want to have their content in AI. There's a different class of companies, and they tend to be companies that are either ad-supported or subscription-supported. And in those cases, that's where the people are feeling like their content is being stolen from them. And in those cases, again different from the majority of companies, those publishers want to have the controls to say, "You can only get my content if you compensate me for it, because you're going to hurt me in terms of advertising revenue and everything else." The efficiency still matters—we still want them to be much more efficient—but then we also want to give them the tools to be able to say, "I want to enforce what the rules are, where you can't have my content unless you pay for it." And then we want to make it as easy as possible for agents to actually pay for that content and be able to have a real exchange of value.
Host 22:11 ↗
Okay, so there's these two buckets: one is the bucket of companies that want their stuff crawled, which is most of them, but then we do get to the news publisher side. And there is a risk here, and we've talked about it on this show a lot, that the incentives for news publishers and information publishers will go to zero if the bots just crawl the information they put on their websites and then present it to them in the chatbot format. I think you've mentioned that people don't really click the footnotes anymore; they just trust what the bot says. And so, let me share again the stats about where things looked last year. Last year you told us that 10 years ago on Google, Google would crawl your page twice and send one click for those two crawls. Six months ago, it was six crawls for one click. Once Google released AI Overviews mid last year, it became 18 crawls for one click. OpenAI had 1,500 crawls for every click that it sent. Anthropic and Claude were about 60,000 crawls for every click that it sent to publishers. Where do those numbers stand today?
Matthew Prince 23:18 ↗
Everything has gotten basically—assuming you're a publisher, from the publisher's perspective, it's gotten twice as bad. Just double every one of those numbers over the last year. Now, from an end user's perspective, I think what you see is actually something quite different, which is it seems like the world has gotten better. The AI systems are smarter, they seem like they're responding better, they're giving you the answers that you need without you having to double-check them. And those two things are actually completely consistent, because it's not like the UI changed, it's not like the behaviors of the AI companies changed; what changed is the results got better. And as the results got better, people click through to the sort of original sources much less. And that I think is indicated by the fact that you've had again a very substantial further drop in that. And that's the trend. Over time, even Google has announced that the era of 10 blue links is going away and that everything is going to look much more like an AI interface. And I think that that's the change. If that's the case, then the business model which has supported the web for the last 30 years, which has been advertising and subscription-based, isn't going to be the business model that supports it for the next 30 years. And we've got to figure out what that is. For sure some of that is going to be deals between the big publishers and the AI companies; we're seeing those increase substantially. But we've got to have something that actually works for everyone. And so one of the things that we're thinking about is how do you have actually micro-payments where, if you're accessing a piece of information, a site can say, "Listen, you can have this, but it's going to cost you a fraction of a penny." And the minimum that's going to cover the infrastructure cost of serving it, but in the case of people who are doing real journalism or academic work or other things, maybe it helps cover the cost of actually supporting the creator that's behind it as well. But it's crazy, the volumes. Here's the interesting technical problem: Visa, largest payments network in the world, handles at peak during the holidays somewhere around 20,000 financial transactions per second. Cloudflare handles about half a billion internet transactions per second. Those are not financial transactions, those are different stakes. But we think that somewhere between 1 and 10% of those will day one be monetizable through agents with a micro-payments transaction, which means that if you're going to build something like this, you've got to build a payments infrastructure that can day one support somewhere around 10 and then be able to scale to over 100 million financial transactions per second in order to be able to support what's there. Now, these are going to be really, really, really small transactions, but they're still financial transactions. You've still got to account for them and do that. That's an incredibly interesting and hard technical problem, and we're looking for both external partners that can solve that, but we're also working on how internally we can
Build that. And those are the sorts of things that you have to have in place in order to then support what is a new business model for the internet. And again, I can't imagine any problem that's more exciting to be working on right now.
Host 26:14 ↗
Yeah, so a year ago basically the solution that you proposed or sort of implemented was a system where if you want to crawl a publisher page, you would have to pay them a micro-payment, and if you didn't pay it, then you'd be blocked. So—

30 more exchanges in this transcript

Sign in free to read the rest of this interview. No card required.

Sign in to read the full transcript

Cite this transcript

APA, MLA, BibTeX
APA

Prince, M. (2026, September 2). Cloudflare CEO: We're Ready To Block Millions of Websites From AI — With Matthew Prince [Interview transcript]. Alex Kantrowitz. CEOInterviews.AI. https://ceointerviews.ai/interview/1298061/

MLA

Matthew Prince. "Cloudflare CEO: We're Ready To Block Millions of Websites From AI — With Matthew Prince." Alex Kantrowitz, 2 Sep. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/1298061/.

BibTeX
@misc{prince2026_1298061,
  author       = {Matthew Prince},
  title        = {Cloudflare CEO: We're Ready To Block Millions of Websites From AI — With Matthew Prince},
  howpublished = {Interview transcript, Alex Kantrowitz. CEOInterviews.AI},
  year         = {2026},
  month        = {sep},
  url          = {https://ceointerviews.ai/interview/1298061/},
  note         = {Speaker-attributed transcript with timestamps}
}