Two technology-policy specialists asked the D.C. Circuit, the federal appeals court hearing the case, on August 4 to toughen the remedy against Google, arguing generative AI cannot check a search monopoly that funded its own AI buildout. Joel L. Thayer, who leads the Digital Progress Institute as its president, and Asad Ramzanali of the Vanderbilt Policy Accelerator filed the 34-page brief. It is an amicus filing, an argument submitted by outside parties rather than a ruling, backing the Department of Justice and state plaintiffs in the consolidated appeal, case No. 26-5023.

The filing targets a specific piece of reasoning. In August 2024, U.S. District Judge Amit Mehta ruled that Google had unlawfully held monopoly power over general search and search text advertising. His September 2025 remedies order rejected a Chrome divestiture and ordered syndication and data-sharing requirements in its place, citing what he called a “healthy dose of judicial humility” toward generative AI’s uncertain trajectory. Google’s notice of appeal followed in January 2026.

The brief’s first argument is financial. It traces a line from Google’s 2011 Project Marvin research effort and the 2013 purchase of DNNresearch, which brought in researchers including Geoffrey Hinton, through the $650 million DeepMind acquisition in 2014 and the 2015 rollout of RankBrain. Amici describe the resources that entrenched Google in search as the direct inputs to its current AI position, not a coincidence of timing.

A second section frames generative AI as a four-layer stack: chips, cloud infrastructure, foundation models, and consumer applications. The brief cites a Menlo Ventures report from December 2025 finding that Anthropic, OpenAI, and Google together held a combined foundation-model API revenue share of 88 percent, with Google’s individual slice at 21 percent as of last December. For the layers beneath, it cites unnamed industry estimates putting Nvidia above 80 percent of AI chip supply and figures showing Amazon, Microsoft, and Google controlling roughly two-thirds of cloud revenue, Google’s own portion sitting at 15 percent.

The application layer is the point where the brief’s argument reaches search marketers most directly. It cites Similarweb data putting AI Overviews on screen for 43.1 percent of all Google search results by May 2026, alongside an estimated 279 million monthly visits to AI Mode, Google’s conversational search experience. That figure is not the only measurement available. PPC Land has reported that Adthena’s June 2026 tracking put AI Overview prevalence at 18 percent of U.S. searches and 23 percent in the U.K., a gap no published methodology has reconciled.

Independent measurement is more consistent on outcomes. Ahrefs’ February 2026 research linked AI Overviews to a click-through-rate decline of 58 percent on pages ranking first, a jump from the 34.5 percent decline the firm had measured a year earlier. An April 2026 randomized study of 1,065 Chrome desktop users found organic click volume dropped 39.8 percent whenever an AI Overview rendered on the results page, while sponsored-click volume held steady.

Amici also argue that entering an adjacent market does not excuse a monopolist holding power in the original one. They point to the Apple relationship as the clearest repeat pattern. Mehta’s remedies order lets Google keep paying Apple roughly $20 billion annually to remain the default in Safari, provided the arrangement stops short of full exclusivity, and in January 2026 Google and Apple confirmed Gemini would power Siri, an arrangement Apple reaffirmed at WWDC26 in June.

No date has been set for oral argument, and the brief carries no binding weight of its own. What it does is name a specific number, Google’s 21 percent of foundation-model API revenue, that the district court did not have when it weighed AI as a future check on search competition. Search teams tracking AI Overview volatility should watch the appellate timeline as closely as any Google product change: the syndication and data-sharing rules that took effect in February 2026 are what stand between the current distribution advantage and a genuinely contested market.

Reporting by Luis Rijo for PPC Land, published August 8, 2026.