AI-powered bots are flooding publisher sites, driving up costs and siphoning audiences. New data reveals a 62.5% surge in bot-driven sessions, with AmazonBot leading the pack.
Publishers are dealing with a new kind of traffic problem: AI bots that bring no revenue but plenty of complications. Over the past year, the share of website visits from AI harvesting bots has jumped by 62.5%, now making up 13% of all web sessions, according to new data from 51Degrees. For content creators and digital media companies, this isn’t just a technical hassle-it’s a direct threat to their business.
In June 2026, Cloudflare reported that automated requests accounted for 57.5% of HTML web traffic on its network, marking the first time machine traffic surpassed human traffic at this scale.
AmazonBot, Amazon’s web crawler, is the most active, responsible for 4.1% of all web visits tracked in the study. Meta-ExternalAgent follows at 2.9%, and Microsoft BingBot at 1.35%. Other AI bots include Anthropic at 0.19%, OpenAI at 0.09%, Perplexity at 0.04%, and Google Gemini at just 0.0003%. Google’s main search bot-which also gathers data for AI-was not included in these numbers, so the real impact is likely higher.
These bots don’t count as valid advertising traffic, so publishers see no benefit in their analytics. 51Degrees notes that while some bots identify themselves, many try to look like human visitors, making them harder to spot and block. The company’s analysis, based on over 3 billion website visits in the year up to May 2026, only includes bots that declare their identity. That leaves out more sophisticated or malicious actors.
Cloudflare reported that in June 2026, training crawlers made up about 52% of all AI crawler requests, up sharply from 22% in spring 2025. This indicates a significant shift toward AI model training rather than traffic that could send users back to publishers.
The effects are immediate. Publishers have to invest in stronger infrastructure to handle the extra bandwidth, while also trying to protect their content from being copied and reused by AI platforms. If nothing changes, the result is lost ad revenue, weaker audience loyalty, and a sense that the open web is being stripped for data.
Some publishers are looking for ways to regain lost ground, such as re-engaging lapsed subscribers-a tactic covered in an earlier report-but these efforts are undermined if AI bots keep eroding the core audience and the value of original content.
This isn’t just a technical fight. As AI bots get more advanced and harder to block, the open web risks becoming a data source for tech giants, while the publishers who create the content pay the price. Whether independent publishing can survive may depend on how much of this silent siphoning the industry is willing to accept.
Regulators are starting to pay closer attention. On September 1, 2026, EU antitrust officials began asking publishers about Google’s AI search opt-out, reflecting concerns that AI-generated summaries can cut referral traffic and revenue for media outlets, according to a Reuters investigation. This scrutiny highlights what’s at stake for both publishers and tech platforms as the situation continues to change.
