According to Help Net Security's coverage of DataDome's AI Traffic Report for Q3 2026, AI agent traffic across DataDome's network grew 41.5% quarter over quarter to nearly 25 billion requests. Meta's crawlers, Meta-ExternalAgent and Meta-WebIndexer, generated 12.4 billion of those requests, making Meta the dominant source of AI agent traffic. The report also highlights Meta's new Muse agent, which produced 29.1 million requests in its first 12 days of full availability, with 99.3% of that traffic coming from cloud and CDN infrastructure.
Agentic browser traffic—browsers driven by AI agents rather than scheduled crawlers—rose sharply in the same period, from 12.5 million requests in August to 58.2 million in September. Even excluding Muse, that traffic more than tripled to 38.7 million requests. DataDome's analysis found that Muse's daily traffic patterns matched user activity, not a crawl, and that 8.9% of its requests reached sensitive endpoints such as login, payment, account creation, and forms. Forms accounted for 81% of that sensitive traffic, receiving about eight times as many requests as login pages.
The report's authors argue that this shift undermines policies that trust residential or ISP-based traffic, since agents like Muse operate from cloud virtual machines. They recommend that organizations validate agent identity, assess intent, and decide deliberately which agents to allow and what they are authorized to do. The single source for this article is Help Net Security's write-up of DataDome's report; no other sources were available for comparison.