Google’s own brief says the “open web” is in rapid decline

Court strategy collides with months of “web is thriving” spin, just as AI Mode creeps toward default.

In a filing late Friday, Google told a federal judge that the “open web is already in rapid decline,” warning that forced divestitures in its ad-tech case would make things worse for publishers. The line appears in Google’s Sept. 5 remedies memo. That’s the tell.

Publicly, Google has said the opposite. In May, CEO Sundar Pichai talked up a bigger, livelier web; senior search leaders echoed that the web is “thriving.” The courtroom argument flips the script, and not by accident.

What changed this week

Two tracks converged. On the policy track, the company is fighting U.S. proposals to split parts of its ad stack, arguing such remedies would harm publishers who still depend on open-web display ads. On the product track, a Google product lead teased that “AI Mode” will soon be the default search experience; the company later dialed that back, but google.com/ai now routes users straight into the chatty interface. The signal is clear enough. Google is shifting from gateway to platform.

The filing tries to reconcile both worlds. It frames the decline around open-web display advertising, not the web writ large. That nuance matters—and doesn’t.

The legal gambit

In court, Google’s message is simple: the market is already moving. Ad budgets are flowing to connected TV, retail media, and AI-driven formats. Forcing a breakup, Google says, would accelerate publishers’ pain rather than restore competition. It’s classic remedies judo—argue that structural change risks unintended damage in a fragile ecosystem.

Critically, the company now distinguishes “the web” from “open-web display.” That’s narrower. But it also mirrors what publishers experience: fewer referral clicks, smaller open-web ad yields, and traffic siphoned by AI summaries. The filing’s logic doubles as a diagnosis.

Publishers’ AI dilemma

Many publishers spent the year locking out AI crawlers. Then they discovered invisibility cuts both ways. As Common Crawl noted, brands that block ccBot can vanish from AI answers while rivals appear; some reversed course to regain visibility. It’s a whipsaw.

Platforms are drawing harder lines, too. Reddit moved to block the Internet Archive’s Wayback Machine from archiving most pages, citing AI firms that mined the archive to dodge platform policies. That decision preserves licensing leverage. It also breaks a public record that journalists and researchers rely on. The tradeoff is stark.

Regulation accelerates the walling-off

The U.K.’s Online Safety Act shows how rules aimed at harm reduction can fragment access. Wikipedia’s operator lost an early court challenge over how regulators might categorize the site. The court urged caution, but the risk is real: identity checks and age-gating requirements could burden nonprofits and small communities that lack compliance budgets. Some sites will comply; others will geoblock; many will simply disappear.

In the U.S., several states have passed age-verification laws for large platforms and adult sites; federal deepfake rules now mandate rapid takedowns. Good intentions, complex enforcement. Fragmentation follows.

The new economics of access

Cloudflare has started normalizing “pay-per-crawl,” letting sites charge AI bots or block them by default. That feels empowering to publishers in the short term. Over time, it favors intermediaries who run the tollbooths, not individual creators. Meanwhile, Cloudflare accused Perplexity of stealth crawling to bypass robots.txt, underscoring the brittle truce between AI firms and content owners. Everyone is building gates. Some will be jumped.

If AI Mode becomes the center of search, the incentives harden. The more Google answers inside Google, the less the web functions as a commons—and the more valuable paid pipes into that answer box become. A handful of big licensing deals can dominate distribution. The long tail gets dimmer.

What Google is really saying

Inside the courtroom, Google is telling the judge the open-web ad economy is shrinking and sensitive to shocks. Outside, it insists the broader web is fine and that AI-infused search still sends “billions of clicks.” These claims can coexist: the total corpus of pages may grow while economic oxygen and discoverability concentrate elsewhere. Both can be true. The consequence is what counts.

The deeper contradiction isn’t about word choice. It’s about control. Google thrives when the web is open enough to index, yet closed enough that answers and transactions happen on its surfaces. Publishers survive when the web remains legible, linkable, and worth visiting. Those goals are diverging.

Where this leaves the web

We are drifting toward three architectures: licensing gardens, pay-to-crawl pipes, and compliance-walled zones. Each de-emphasizes the open link economy that funded independent sites for two decades. Each rewards scale, lawyers, and platform gravity over open protocols.

The court will decide remedies. Markets and product choices will decide everything else. Unless incentives change, the open-web share of attention—and money—keeps sliding. That’s the part of Google’s sentence that matters.

Why this matters

• Consolidation accelerates when distribution shifts inside platforms and compliance costs rise, squeezing small and independent sites.
• AI visibility becomes a paid advantage; those who can’t or won’t license access risk vanishing from where users now get answers.