Some of the web's biggest publishers are preparing to block Google entirely

On Sept. 15, 2026, Cloudflare will change a default that has shaped the open web for a decade. The company sits between most websites and the public internet, routing traffic for a large share of the web. Starting that day, it will begin blocking "multi-purpose" crawlers, the bots that scrape pages for both search indexing and AI training, unless site owners opt them in. For the first time, a meaningful slice of publishers can unplug from Google with a single toggle instead of a lawsuit.

The publisher is "blocking other crawlers" already and could turn off Google's access too if search traffic keeps falling. USA Today Co. is separately suing Google over an alleged ad-tech monopoly, so the AI-search question is the second front in a longer fight.

The numbers behind the posturing come from Semrush, the third-party measurement firm, and were relayed by Nieman Lab from a Wall Street Journal report. Between June 2025 and June 2026, more than 40 percent of traffic to some publications disappeared. The Guardian and BBC went the other way and gained readers in the same window. The collapse is not universal, but it is concentrated, and the publishers hit hardest are the ones whose pages Google increasingly answers with an AI summary rather than a link.

Cloudflare is not a neutral referee in this fight. Its core business includes bot management, and the company's chief strategy officer, Stephanie Cohen, told Adweek the new default was a way to make content "discoverable without surrendering it to free AI training." Bots now account for more than half of all web traffic, Cloudflare says, and the company is offering publishers the cheapest opt-out in the market.

Reddit is reportedly reevaluating its roughly $60 million per year licensing deal with Google, a sum the company has not committed to renewing. Politico and Reuters are weighing registration walls and crawler limits of their own. Penske Media sued Google over AI summaries in 2025. People Inc. CEO Neil Vogel has publicly opposed AI scraping, though his most recent comments were partly truncated in the Nieman Lab relay.

For most publishers, the practical question is what an opt-out actually buys them. Google's crawler is one of a handful of bots that decide whether a page appears in search results, and search remains the largest free referral channel on the public web. A block does not just remove a publisher from AI summaries; it removes the publisher from Google search itself. The Sept. 15 default only changes the cost of trying, not the cost of losing that channel.

The phrase for what publishers are bracing for was coined in 2024 by The Verge editor-in-chief Nilay Patel: "Google Zero," the moment when search stops sending traffic off-platform and starts answering questions itself. Two years on, it has moved from a warning to a budget line.

The first batch of opt-outs lands in mid-September. By year-end, the industry will know whether the lever that moved from the courtroom to the data center actually changed the answer.