Fastly has released research suggesting AI-generated traffic is becoming a distinct layer of internet activity, with implications for how organisations manage automated requests, protect infrastructure, and make decisions about content access.
Based on analysis of traffic across Fastly’s global network, the company said AI requests grew by about 30% between January and May 2026—around 6.5 times faster than human traffic over the same period. Fastly argued the shift is not only volume growth, but changes in how AI systems interact with digital infrastructure.
The company said organisations are increasingly weighing business and operational trade-offs in responding to automated traffic, beyond traditional bot-blocking. Fastly’s research described “machine traffic” as including AI crawlers, AI fetchers, bots, agents, and API-driven systems.
“AI traffic is fundamentally changing how the internet operates,” said Artur Bergman, Founder and Chief Technology Officer at Fastly. “Businesses are moving beyond a world where humans are the primary users of digital experiences. The challenge is no longer simply blocking bots, it’s understanding which machine interactions should be accelerated, managed, challenged, or stopped.”
Fastly identified two categories of AI traffic in its dataset: AI crawlers, which systematically gather information to build and update AI models, and AI fetchers, which retrieve information in response to specific user requests via AI assistants and agentic applications.
Using May 2026 data, Fastly said these workloads place different demands on infrastructure. The company reported that 51% of AI requests required origin access, compared with less than 9% of human requests. It also said “Claude-related traffic” increased by more than 555% compared with its January 2026 baseline.
The research also pointed to variation in how large organisations respond to AI traffic. Fastly described one example where a company introduced a hard block following a sudden spike in AI fetcher traffic, and another where an organisation did not block AI agents, followed by increased fetcher volume over several months.
Fastly said the findings indicate AI traffic management is expanding from a security and infrastructure issue into a broader strategic consideration, where visibility and control over how AI systems interact with content, applications, and APIs are becoming more important.
