Recently a headline number stopped me in my tracks: automated traffic now officially exceeds human traffic on the internet. Not in some niche corner. Across the entire Internet.

That is not a prediction about 2028. That is what happened by end of 2025.

The data is clear

HUMAN Security analyzed over 1 quadrillion digital interactions in 2025. Their findings: automated traffic grew 23.5% year over year, while human traffic grew just 3.1%. That is an 8x difference in growth rate. AI driven traffic specifically surged 187% over the course of the year, nearly tripling in 12 months.

Imperva’s 2025 Bad Bot Report, covering 2024 data, had already flagged the crossover: bots accounted for 51% of all web traffic that year. The first time in a decade that machines outpaced humans. By the end of 2025, Cloudflare’s own numbers showed the split at roughly 53% bot versus 47% human on HTML requests across their network, which serves about 1 in 5 websites globally.

Cloudflare CEO Matthew Prince predicted at SXSW in March 2026 that bot traffic would surpass human traffic by 2027. Other sources suggest that was conservative. It has already happened.

Three types of AI traffic are reshaping the web

Not all AI traffic is equal.

Training crawlers still make up the bulk at roughly 67% of all AI driven traffic. These are the bots that scrape the web to feed large language models.

Real time AI scrapers are the faster growing category. These are the bots that fetch live data when you ask ChatGPT a question or when an AI product checks a price for you.

Agentic AI is the smallest category today at about 1.7% of AI traffic, but it grows at insane rates. These are autonomous agents that browse, fill forms, manage accounts, and complete purchases without human intervention.

The shift from passive crawling to active participation is the real story. We are moving from bots that read the internet to bots that use the internet.

The 100x multiplier problem

Matthew Prince shared a powerful example. When a human shops for a digital camera, they might visit 5 websites. An AI agent performing the same task visits 500. That is a 100x multiplier in traffic load for a single query.

AI driven growth is continuous. Prince put it plainly: internet traffic keeps growing and there is nothing on the horizon that will slow it down.

Global internet traffic grew 19% in 2025 according to Cloudflare, up from 17% the year before. AI bots are a primary driver of that acceleration. And the nature of the traffic is changing. Training crawlers can be throttled and scheduled. Real time scrapers and autonomous agents demand instant responses. That means the load is shifting from batch processable to latency sensitive, which puts very different demands on infrastructure.

What this means for enterprise connectivity

If you run enterprise networks, this should change how you think about capacity planning.

The traffic hitting your infrastructure is no longer just your employees and your customers. It includes AI crawlers indexing your content, AI scrapers pulling real time data from your services, and increasingly, autonomous agents interacting with your applications (mostly by your own users!). That traffic is growing at multiples of overall internet growth.

Bandwidth provisioning that accounts for 15 to 20% annual growth based on historical human usage patterns is no longer sufficient. AI driven traffic can spike unpredictably and new applications or features can instantly drive large traffic volumes. It consumes server resources differently than human browsing, and it is only accelerating.

The internet yesterday is not the internet we will tomorrow

We built internet infrastructure assuming a human sits on the other side of every connection. That assumption is changing rapidly. The internet has crossed a structural inflection point where machines are the dominant users, and their share is growing faster than anything we have seen since the early days of the web.

For anyone responsible for enterprise connectivity: the time to plan for this is now. Your networks need to be ready.

Sources