For years, the internet was built around humans and their searches. That is beginning to change.
According to Cloudflare’s Radar – a free, publicly accessible internet observability platform – bots now account for 57.5 per cent of global web requests, compared with 42.5 per cent generated by humans. Cloudflare co-founder and chief executive officer Matthew Prince on June 3, 2026 said this marks the first time in the internet’s history that bots have overtaken human traffic — a milestone he had expected to arrive in 2027.
At first glance, the figure may sound like a story about spam or malicious automation. But Cloudflare’s data points to something broader. Not all bots are harmful or invisible scripts. A growing share now performs useful, structured tasks across the internet, including retrieving information, comparing options and assisting users through AI-powered systems.
The shift signals a structural change in how the web is being used. Increasingly, bots are no longer just background infrastructure. They are becoming active participants in how information is discovered and consumed.
Not all bots are the same
Bots are not new to the internet.
For years, search engines have relied on crawlers to index websites. Monitoring tools have checked uptime, while security systems have scanned infrastructure. These automated processes have always accounted for a significant share of internet traffic. What is changing today is the composition of that traffic.
According to content delivery network provider Fastly, bots broadly fall into different categories based on their functions. To understand the shift, it helps to distinguish between three increasingly important groups:
-
AI crawlers: These systems systematically scan websites and collect information that can later be used for indexing or training large language models. Their behaviour resembles traditional search-engine crawlers, although their purpose now extends beyond search. -
AI fetchers: Fastly describes these as systems that retrieve information in response to specific user requests. When someone asks an AI assistant a question requiring current information, these bots may visit websites to collect relevant data such as prices, availability or facts before generating a response. -
AI agents: Cloudflare points to these as emerging systems capable of performing multi-step tasks on behalf of users, such as comparing flights, checking product prices, summarising information across websites or completing basic online actions.
Together, these categories are creating a rapidly growing layer of automated activity across the internet.
Why bot traffic is rising
The growth in bot traffic is closely linked to the rise of AI assistants that increasingly interact with the web.
This represents a significant shift in how information is accessed online.
Traditionally, search engines functioned as discovery tools. Users entered a query, received a list of links and visited websites to find answers. AI-powered systems increasingly function as answer engines. Instead of directing users to multiple pages, they gather information in the background, process it and return a consolidated response.
Fastly’s data illustrates how quickly this layer is expanding. Between January and May 2026, AI-related bot requests on its network rose around 30 per cent — roughly 6.5 times faster than human traffic over the same period. The company also reported that Claude-related traffic increased more than 555 per cent compared with its January baseline.
This does not mean AI bots dominate the internet on their own. Cloudflare’s figure refers to overall bot activity, which includes many forms of automation. However, AI-related bots are increasingly becoming an important driver of growth within that broader category.
Bots behave differently from humans
One reason bot traffic matters is that bots do not behave like human users. Human activity follows predictable rhythms. Usage rises during waking hours, follows work schedules and fluctuates across weekdays and weekends. Bots operate continuously. Fastly found that crawler activity remains largely steady throughout the day. These systems do not sleep, pause or wait for peak usage periods. They continuously scan, collect and synchronise information across the web.
Bots also interact with infrastructure differently.
According to Fastly, fewer than 9 per cent of human requests on its network require direct access to origin servers, where website data is stored. By contrast, more than 51 per cent of bot-related requests require direct origin access.
This is because bots often seek fresh, real-time information, including updated prices, inventory levels, breaking news and changing documentation, which cannot always be served from cached copies.
For businesses, this means bot traffic is not just growing in volume. It is also increasing infrastructure demand.
Why publishers are paying attention
The rise of bots that retrieve and synthesise information is reshaping the relationship between publishers and platforms. For years, publishers relied on search engines for distribution. Search engines crawled websites, indexed pages and directed users back to original sources. Traffic and advertising revenue flowed back to publishers in exchange for discoverability.
That model is now under pressure.
AI-powered systems such as Google’s AI Overviews, ChatGPT and Perplexity increasingly combine information from multiple sources and present it directly to users. Instead of visiting individual websites, users often receive complete answers within a single interface.
This has contributed to what many describe as a “zero-click internet”, where content is consumed without users visiting the original source.
The issue has become significant enough to attract regulatory attention.
The UK’s Competition and Markets Authority recently instructed Google to provide publishers with greater control over whether their content is used in AI-powered search features. Google has since begun testing controls that allow publishers to opt out of certain AI experiences while remaining visible in traditional search results.
At the centre of the debate is a fundamental question: if bots increasingly retrieve and summarise content on behalf of users, how should the value created by publishers be distributed?
Ecommerce is facing a similar shift
The impact of rising bot traffic is not limited to publishing. In ecommerce, bots are increasingly involved in how products are discovered and compared. Cloudflare notes that many AI-powered bots already perform tasks such as reading product pages, checking prices, comparing options and summarising information for users. At the same time, technology companies are building systems that formalise this behaviour.
OpenAI has expanded shopping-related capabilities within ChatGPT, allowing users to conduct product research and comparisons through conversational queries.
Google is also investing in what it describes as agentic commerce, where AI systems actively assist users in evaluating products and services. The company outlined aspects of this vision during its recent Google I/O conference.
The implication is a shift in how shopping begins.
Instead of manually browsing multiple websites, users may increasingly rely on bots to handle the initial stages of discovery and comparison. Retailers may therefore need to optimise not only for human shoppers but also for the AI systems influencing what users ultimately see and choose.
For years, retailers competed for visibility on search engines and marketplaces. In a more agent-driven internet, they may also need to compete for visibility within AI systems that recommend products on behalf of users.
The customer may still be human, but the browsing, comparison and evaluation process could increasingly be performed by machines.
A new phase of the internet
The significance of Cloudflare’s data is not simply that bots now outnumber humans in web requests. It is what those bots are doing.
Some continue to perform traditional functions such as crawling and monitoring. Others are increasingly retrieving real-time information, answering questions and assisting users with tasks across the web.
Humans remain central to the internet economy. People still create content, make decisions, consume entertainment and drive demand.
But the path between users and websites is changing.
For decades, the internet was primarily a space where people visited websites directly to find information. Increasingly, that process is being mediated by bots that browse, retrieve and summarise information on their behalf.
The internet is not becoming less human. But it is becoming far less direct.
