Fastly says AI requests grew 6.5 times faster than human traffic

AI-generated traffic is becoming a distinct layer of internet activity, according to new research from Fastly, as organisations decide how automated systems should interact with their content, applications and APIs.

The edge cloud platform analysed traffic across its global network and found that AI requests grew approximately 30% between January and May 2026. That was about 6.5 times faster than human traffic during the same period.

Fastly’s findings point to a shift in how businesses manage machine traffic. Blocking bad bots remains part of the task, but organisations also need to decide which AI systems should be allowed, limited, challenged or stopped, especially as automated requests affect infrastructure, content visibility and digital distribution.

AI crawlers and fetchers create different demands

Fastly identified two main categories of AI traffic: AI crawlers and AI fetchers.

AI crawlers systematically gather information from the web to build and update AI models. AI fetchers retrieve information in response to user requests through AI assistants and emerging agentic applications.

The distinction is important because the two categories behave differently and serve different purposes. Crawlers are linked to model training and content collection, while fetchers are tied more directly to live user activity, including answering questions, comparing options, validating facts and completing tasks.

The research also found that machine traffic now includes AI crawlers, AI fetchers, bots, agents and API-driven systems. Fastly said organisations are weighing content protection, visibility, customer acquisition and digital distribution as they decide how to manage these interactions.

“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.”

AI requests are reaching origin systems more often

Fastly’s May 2026 data found that more than half of AI requests, at 51%, required origin access. That compares with less than 9% for human requests.

The company said this shows that AI workloads place different demands on digital infrastructure. As AI systems retrieve, process and act on information, organisations need more visibility into which automated systems are interacting with their digital properties and how those systems behave.

Fastly also observed rapid growth in traffic linked to specific AI systems. Claude-related traffic increased by more than 555% compared with its January 2026 baseline.

Businesses are taking different approaches to AI access

Fastly’s analysis described a split in how companies respond to AI traffic. In one case, a large company introduced a hard block after a sudden spike in AI fetcher traffic, most likely to maintain content authority. In another case, a large company chose not to block AI agents, which led to higher fetcher volume over several months and potentially greater visibility through AI-powered services.

The examples show how machine traffic decisions can affect both infrastructure management and customer engagement. Restricting AI access may help protect content or control automated demand. Allowing selected AI systems may improve visibility as users rely more on AI assistants to discover, compare and act on information.

Fastly also highlighted that an effective machine traffic strategy requires visibility into which AI systems are interacting with digital properties, context around whether those systems create business value, and precision in responding differently based on intent and impact.