Automated traffic on the internet already outpaces human traffic, and Cloudflare, one of the largest infrastructures on the network, wants to change the rules of the game. Its plan: turn content crawling by artificial intelligence bots into a paid service, rather than a free resource.
The proposal is gaining relevance now because the company's own internal data confirms that the balance between human and automated visitors has already broken down, and it happened sooner than expected.
Bot Traffic Already Surpasses Human Traffic
In November 2025, during the European Web Summit, Cloudflare CEO Matthew Prince presented an internal projection placing the crossover point between automated and human traffic in the second half of 2027.
That forecast didn't hold. By March 2026 the date had moved up to the first half of that year. By May 2026, according to the company's own data, the crossover had already occurred: bots, crawlers and AI agents were generating more requests on the network than people were.
Cloudflare manages a significant share of global traffic, giving it direct visibility into these flows. Prince has gone further with a forward-looking projection: if the current curve holds, automated traffic could end up being a thousand times greater than human traffic within five years, not because human visits decrease, but because of the speed at which AI agents are expanding.
A Structural Position Between the Web and the Bots
Cloudflare operates infrastructure in more than 350 cities, with thousands of distributed data centers. That network places it literally in the middle of the traffic reaching millions of sites, both from human visitors and automated systems.
That position allows it to offer website owners three paths when it comes to artificial intelligence bots: block them completely, let them operate freely, or require payment for access to the content.
It's this last option that Prince is developing as a business model. In practice, a media outlet, a digital library or an online store could use Cloudflare's infrastructure to make an OpenAI or Anthropic agent pay upfront before crawling its data.
The Advertising Model That Doesn’t Work With Bots
Prince describes the last three decades of the internet as an era dominated by advertising, with Google as the main architect of that system through acquisitions like DoubleClick and tools like Google Analytics.
That model depends on the visitor to a page being a person capable of seeing an ad. When the visitor is a bot extracting text to train a language model or power an AI assistant, there's no ad impression, no click, and no revenue for whoever produced the content.
The result is a structural problem: the more useful a website is, the more it gets crawled by artificial intelligence systems, consuming more resources without generating any economic return for the publisher. By sitting in the middle of that flow, Cloudflare positions itself as the player capable of intervening in that dynamic and charging a toll.
Internal Layoffs and a Parallel Bet on AI
The transformation Cloudflare is proposing for the rest of the web is also reflected within the company itself. In early 2026, Cloudflare laid off more than a thousand employees, representing 20% of its workforce.
Prince did not present these layoffs as a routine restructuring. He signed an opinion piece in The Wall Street Journal about how he decides which Cloudflare employees to replace with artificial intelligence, a piece that openly laid out the reasoning behind the decision.
The contradiction is striking: Cloudflare is proposing to charge AI companies for consuming web content while using that same technology to shrink its own human workforce.
A Tension That Remains Largely Unresolved
During an interview, Nilay Patel, editor-in-chief of The Verge, directly asked Prince whether the outcomes he describes are actually good for the future of the web. According to Patel himself, Prince's answer stayed with him even after the conversation ended.
The underlying tension is clear. If artificial intelligence bots pay to access content, publishers could receive revenue they don't currently have. But that same model could concentrate access to information in the hands of the AI companies with the greatest financial resources, leaving smaller players or non-commercial uses out in the cold.
Furthermore, if automated traffic ends up being a thousand times greater than human traffic, someone will have to bear the cost of the infrastructure needed to sustain it. Prince argues that this cost must necessarily fall on those generating that traffic: the artificial intelligence companies themselves.
What This Shift Means for Publishers and Users
For those who publish content on the web, whether a blog, an online store or a media outlet, Cloudflare already offers the option to block or attempt to monetize bot access. That decision stopped being theoretical in May 2026, when automated traffic surpassed human traffic, meaning more than half of the visits to many sites no longer come from real people.
For those who use artificial intelligence assistants, the cost of these services could rise if the companies developing them start paying publishers for access to their data. The expense of training and running these systems doesn't disappear, it simply shifts elsewhere.
And for the tech sector as a whole, Cloudflare's case illustrates a concrete advantage: those who control the network's infrastructure are better positioned to charge a toll on any new business model that emerges, regardless of who produces the content or who ultimately consumes it.
Frequently Asked Questions
What is Cloudflare and why does it carry so much weight on the internet?
Cloudflare is a company that manages network infrastructure for millions of websites, with a presence in more than 350 cities. Its position allows it to see and control a large share of the traffic flowing through the internet, both human and automated.
When did bot traffic surpass human traffic?
According to Cloudflare's internal data, the crossover happened in May 2026. The company had initially projected this date for the second half of 2027, but moved it up in successive revisions.
How does Cloudflare intend to charge AI bots?
The company is building infrastructure so that website owners can require payment from artificial intelligence bots before allowing them to crawl their content, instead of blocking them or letting them access it for free.
How does this relate to the layoffs at Cloudflare?
In early 2026, Cloudflare laid off more than a thousand employees, 20% of its workforce, in a process its CEO, Matthew Prince, explicitly linked to replacing human tasks with artificial intelligence.
Who is Matthew Prince?
Matthew Prince is the CEO of Cloudflare and the person who has publicly led this proposal to charge AI bots for accessing web content, as well as justifying the company's internal staff cuts.
Will this change affect the price of AI assistants?
It's possible. If artificial intelligence companies start paying publishers to crawl their data, that additional cost could be passed on, at least partially, to the price of AI services for end users.
