Cloudflare's Bet: Building a New Business Model Before the Bots Take Over the Web

Picture the internet as a public square. For thirty years, that square was mostly filled with people: reading, clicking, buying, arguing in comment sections. Advertisers paid to be near that foot traffic, and that money kept the lights on for publishers, bloggers, and small businesses alike. Now imagine the square filling up with something else entirely. Automated visitors. Scrapers. AI agents running errands for people who never show up themselves. According to Cloudflare, that shift has already happened, and it is accelerating faster than almost anyone predicted.

In June, Cloudflare found that bots now make up more than half of all internet traffic. That is not a fringe statistic from an obscure corner of the web. Cloudflare sits in front of a huge share of websites worldwide, acting as a kind of security gate and traffic cop between sites and everyone trying to reach them, whether human or machine. When Cloudflare says something about the state of the internet's plumbing, it is speaking from an unusually well-informed vantage point.

Cloudflare CEO Matthew Prince laid out just how fast this is moving in a recent conversation with The Verge's Nilay Patel on the Decoder podcast. Last November, at the Web Summit conference in Europe, Prince was asked when automated traffic might overtake human traffic online. Cloudflare's data pointed to the second half of 2027. By this past March, at South by Southwest, the updated estimate had moved up to the first half of 2027. Then, in May, Cloudflare's own team told Prince the crossover had already happened. Automated traffic had passed human traffic months ahead of even the most recent forecast.

Extrapolate that trend, and Prince's team believes automated traffic could reach a thousand times human traffic within five years. Human usage isn't expected to shrink. It's the sheer scale of everything else, agents, crawlers, bots doing tasks on people's behalf, that is projected to explode.

That creates a very practical problem. Someone has to pay for the servers, the bandwidth, and the infrastructure required to handle that much traffic. For three decades, advertising covered those costs. But bots don't click on banner ads. They don't respond to a clever brand campaign or a well-placed logo. The entire financial engine that built the modern web, the one Google spent 28 years perfecting through tools like DoubleClick and Google Analytics, simply doesn't function when the audience is mostly machines.

Who Pays When Bots Do the Browsing

This is where Cloudflare is trying to insert itself into the internet's future, not just its present. Because the company sits between websites and the bots trying to access them, it can give site owners a menu of choices: block AI crawlers entirely, allow them freely, or permit access only to bots willing to pay for it. Earlier this year, Cloudflare made blocking AI crawlers the default setting for websites using its service, a significant shift that effectively forces AI companies to negotiate rather than simply help themselves to content. Patreon has already partnered with Cloudflare specifically to keep AI crawlers away from creator content.

That protective instinct isn't happening in a vacuum. Publishers and rights holders have been fighting AI companies in court over exactly this kind of scraping. Sony and Warner Music both sued Anthropic this year in the latest round of AI copyright litigation. The New York Times has its own long-running case against OpenAI, one that recently took an unusual turn when the Trump administration filed support on OpenAI's side. These lawsuits and Cloudflare's infrastructure-level gatekeeping are really two responses to the same underlying problem: AI systems have been trained and operated on other people's work without a clear, consistent system for compensating the people who made it.

Prince's vision goes beyond simple blocking, though. He wants Cloudflare to help build genuine payment infrastructure, a way for content creators to get paid when AI systems use their work, similar to how advertising once funneled money back to publishers. It's an ambitious goal, and a necessary one if the web's current creative and journalistic ecosystem is going to survive the shift he's describing. If bots outnumber humans a thousand to one and none of them are paying for what they consume, the economic foundation that supports websites, writers, and small publishers could simply erode.

Prince argues that AI companies actually need three things to succeed: skilled researchers, computing chips, and, implicitly, access to the content that trains and informs their systems. He noted that talent shortages and chip shortages are already easing as universities restart AI programs and chipmakers beyond Nvidia, including AMD and Qualcomm, ramp up production. The scarcity that remains, and the one Cloudflare is positioning itself around, is access to quality content and a fair mechanism for paying for it.

There's a human dimension to this that goes beyond corporate strategy, too. Cloudflare itself laid off more than a thousand employees, roughly 20 percent of its workforce, earlier this year. Prince then wrote an op-ed in the Wall Street Journal explaining his reasoning under the blunt headline "How I choose which Cloudflare employees to replace with AI." It's a stark reminder that the same technological wave reshaping advertising and content economics is also reshaping who gets to keep their job, even inside the companies building the infrastructure for that transition.

Why It Matters

The stakes here reach well past Cloudflare's balance sheet. If AI traffic keeps compounding while the advertising model that funded the open web keeps eroding, the practical result could be a thinner, more paywalled internet, one where small publishers and independent creators lose the revenue that once let them exist at all. Cloudflare's attempt to build new payment rails between AI companies and content creators is essentially a bet that the web's economics can be rebuilt before that erosion becomes irreversible. Whether that bet pays off will shape not just who profits from the internet's next era, but whether the people who actually make the web worth visiting can afford to keep doing it.