The Web Is No Longer Ours: When Bots Overtook Humans on the Internet

For decades, the internet was conceived, built, and optimized for us. For our fingers scrolling across a screen, for our eyes scanning for information, for our minds weighing a purchase or choosing a restaurant. Yet something has changed, quietly, beneath the surface of servers and data centers scattered across the world. And it arrived sooner than anyone had anticipated.

According to data published by Cloudflare, the American company that protects and monitors approximately 20% of global web traffic, today nearly 60% of requests made to web pages do not come from human beings, but from automated software: crawlers, scrapers, and above all, artificial intelligence agents. It is the first time in the history of the internet that bots have overtaken real, living users. Matthew Prince, CEO of Cloudflare, had estimated this moment would arrive around 2027. It arrived in the summer of 2025, more than a year ahead of even the most optimistic projections.

The reason behind this acceleration has a precise name: AI agents. These autonomous systems do not simply answer a question or generate text on demand. They navigate the web independently, reading pages, comparing prices, collecting data, filling out forms, booking services, and making operational decisions without any human intervention required. If a human consumer visits five websites before purchasing a product, an AI agent can visit thousands in a matter of seconds, analyzing every detail with a precision and speed no human could match. The traffic generated by these systems is exponential, and the curve shows no sign of slowing down.

This scenario brings to mind a theory that, until a few years ago, was dismissed as a fantasy from underground forums: the Dead Internet Theory. Emerging around 2021 in spaces like 4chan and Reddit, this hypothesis argued that the internet was progressively becoming a place dominated by automated content, bots, and algorithms interacting with one another, gradually hollowing out the network of genuine human presence. At the time, it was labeled a conspiracy theory, filed away among the paranoid fantasies of those who see machinations everywhere. Yet today, in light of Cloudflare's data, that theory appears far less outlandish and considerably more prophetic than anyone gave it credit for.

The fundamental difference from the original narrative is that this is not a conspiracy orchestrated by anyone to manipulate users. It is the natural and, in part, inevitable result of technological evolution. Companies adopt AI agents to automate processes, search engines send crawlers to index content, recommendation systems generate automatic requests. The web is becoming an infrastructure that machines use to talk to other machines, with human beings increasingly becoming spectators of a conversation they never started.

The question that emerges, and that no one can yet answer with any certainty, is what all of this means for the future of the network. The internet was designed to connect people, to democratize information, to give a voice to anyone with a connection. If its primary users become artificial intelligences, the rules of the game change entirely: the economic models built on advertising shift, the criteria by which content is produced and distributed transform, and even the very meaning of publishing something online is called into question. Are we perhaps witnessing the birth of a parallel internet, designed not for human eyes but for the algorithms that inhabit it? The data say yes. And the change is already well underway.

Source: Cloudflare Radar — cloudflare.com/radar — data updated as of June 2026. Statements by Matthew Prince, CEO of Cloudflare.