Date: June 18, 2026 ｜ Source: Cloudflare Radar, BlockTempo
Cloudflare Radar's latest data reveals a stunning fact: 34.1% of global HTTP traffic comes from bots. In other words, when your website shows "1,000 active visitors today," over 300 of them are likely not real people.
Cloudflare CEO Matthew Prince has made an even bolder prediction: at the current pace, bot traffic will officially surpass human traffic as early as 2027. This means we are standing at a watershed — the user structure of the internet is being redefined by AI infrastructure demands.
In the breakdown of all bot traffic, AI crawlers account for approximately 22%, making them the second-largest category behind only search engine spiders. These crawlers are dedicated to scraping web content for AI companies, used for model training or real-time user query responses.
In terms of market share distribution, Meta-ExternalAgent leads at about 16.3%, surpassing GPTBot; ClaudeBot is at roughly 11.6%; and Applebot surged 140% in a single month (from 2.97% to 7.15%), signaling that Apple Intelligence is dramatically ramping up its data collection efforts.
More notably, among all AI crawler traffic, training-type crawlers account for 51.5% — meaning more than half of AI crawlers are stockpiling data for future model versions rather than responding to users in real time. In other words, your website server's bandwidth and resources are being used for free by major AI companies to train their products.
For content platforms and enterprises, this means your article might have been "read" 1,000 times, but 800 of those were bots — and bots do not watch ads. The traffic numbers look impressive, but revenue may actually decline. Meanwhile, when 94% of login attempts are bots, traditional manual review logic has completely broken down.
Do you want to be the sucker paying for that 34% of useless traffic? Your server bandwidth, CPU, and memory are being freely consumed by AI giants' crawlers — while you still have to pay for them. Professional ops isn't just about "keeping services online" — it's about precisely identifying and blocking these useless bot flows so your resources go to real users.