Cloudflare’s new policy pushes AI companies to pay for publishers’ content

Cloudflare announced a new policy requiring AI companies to distinguish between web crawlers used for traditional search indexing and those used for AI training or agentic tasks by September 15. Publishers on Cloudflare's network — which covers a significant share of the web — would by default block non-compliant AI crawlers. The move effectively deputizes Cloudflare as an enforcement arm for content licensing disputes between publishers and AI firms.

The policy touches the entire AI training stack: companies like OpenAI, Google DeepMind, Anthropic, and others that scrape the web for training data would need to update their crawler infrastructure and potentially negotiate directly with publishers via Cloudflare's platform. For NET, this could evolve into a toll-booth revenue model — charging AI companies for verified, policy-compliant access to publisher content.

For Cloudflare itself, the bull case rests on a new monetization wedge in a company already growing revenue 29.8% YoY with 74.5% gross margins. If AI companies pay for crawler access or publishers pay for enforcement tools, it adds a high-margin revenue stream to a business still running at -4.7% net margin. The September 15 deadline is a hard catalyst date to watch.

The bear case is that this is a PR and positioning play more than a near-term revenue driver. AI companies could route around Cloudflare, build direct publisher relationships, or challenge the policy. NET's current -$0.29 diluted EPS means the market is already pricing in future profitability — adding an unproven revenue line doesn't materially change the near-term financial picture.

The key question heading into the September deadline is whether Cloudflare converts this policy into a product with measurable bookings, or whether it remains a brand-building exercise that doesn't show up in guidance revisions.