Antitrust Law Daily Wrap Up, ANTITRUST—S.D.N.Y.: Court greenlights modified AdX Publisher Class in Google Digital Advertising Antitrust case, (Dec 15, 2025)

By Martin A. Steinberg, J.D.

The court rejected certification of the AdSense Publisher Class and the Advertiser Class, holding that the proposed representatives’ claims were atypical and inadequate and that individualized issues defeated class treatment.

The federal district court in New York City granted in part and denied in part motions for class certification in the Google Digital Advertising Antitrust MDL, holding that publisher plaintiffs may proceed with a narrowed AdX Publisher Class but rejecting certification of both the AdSense Publisher Class and the proposed Advertiser Class. The court found that common issues predominate as to AdX publishers’ tying and monopolization claims and that the plaintiffs’ expert offered a sufficiently reliable, classwide damages methodology, while excluding instream video ad sales from the class definition. By contrast, the court denied certification of the AdSense class because its sole proposed representative was atypical and inadequate, and denied certification of the advertiser class because the representative’s claims were temporally and factually misaligned with the alleged conduct and would require individualized inquiries that defeated predominance (In Re: Google Digital Advertising Antitrust Litigation , No. 1:21-cv-07034-PKC (S.D.N.Y. Dec. 12, 2025)).

Background. Publishers sell display advertising impressions using ad servers and real-time auctions conducted on ad exchanges, while advertisers purchase those impressions through ad-buying tools. Google offers a suite of products tailored to customer size and sophistication, including DFP and AdX for larger publishers, AdSense for smaller publishers, and Google Ads or DV360 for advertisers.

The publisher plaintiffs allege that Google violated Sections 1 and 2 of the Sherman Act by coercively tying access to its dominant AdX ad exchange to the use of its DFP ad server, and, for smaller publishers, by bundling AdSense server and auction services, thereby entrenching monopoly power in the markets for publisher ad servers and ad exchanges. Separately, the proposed advertiser plaintiff asserts claims of monopolization and attempted monopolization based on alleged auction-manipulation practices, including Unified Pricing Rules and Project Bernanke.

While this MDL was pending, the DOJ and several states pursued a parallel enforcement action against Google in the Eastern District of Virginia based on substantially the same factual allegations. Following the trial, the Virginia court found that ad servers and ad exchanges are distinct product markets and that Google willfully acquired and maintained monopoly power in both through products such as DFP and AdX, including by unlawfully tying publisher access to AdX to the use of the DFP ad server.

Relying on those determinations, the publisher plaintiffs and Hanson sought partial summary judgment in this MDL, arguing that nonmutual offensive issue preclusion barred Google from relitigating issues already resolved in the Virginia action. The court largely agreed, holding that Google is precluded from disputing the separate-market findings, its monopoly power, and willful maintenance of that power, and the unlawfulness of its tying of AdX access to use of the DFP ad server.

AdX Class and the AdSense Class. The publishers proposed two subclasses: an AdX Publisher Class of U.S. publishers that paid Google in fees or reduced revenue for AdX ad-exchange services between December 2016 and March 2024, seeking damages only for AdX sales, and an AdSense Publisher Class of U.S. publishers that paid Google for AdSense package services during the same period, seeking damages limited to AdSense sales.

The Publishers estimate that the AdX Class consists of approximately 4,986 members, with 99.88% having suffered injury and aggregate damages of about $1.717 billion before trebling, and that AdX held roughly a 62.65% to 66.63% share of the advanced auction-platform submarket during the class period. By comparison, the proposed AdSense Class is estimated to include about 712,807 members, 97% of whom were injured, with total damages of approximately $530.8 million, and AdSense is alleged to have held an almost 100% share of the basic auction-platform submarket during the relevant period.

Motion to certify the AdX Class. The court concluded that the proposed AdX Publisher Class satisfies all of Rule 23(a)’s prerequisites. Numerosity was readily established, given the nearly 5,000 putative class members, and commonality was met because the claims present shared questions concerning Google’s alleged tying conduct, monopolization, and the classwide impact of supracompetitive AdX take rates. Typicality was likewise satisfied, as the named plaintiffs, Genius and The Nation, were subject to the same challenged conduct and advanced the same legal theories as the absent class members.

The court further determined that the named plaintiffs are adequate representatives, rejecting Google’s contention that indemnification provisions involving multi-customer management firms (MCM) create disabling conflicts of interest. The court explained that such provisions would likely be unenforceable to the extent they would require parties injured by Google’s alleged antitrust violations to indemnify Google for its own unlawful conduct. The court also found that class counsel have demonstrated the ability to prosecute the action vigorously and competently and therefore appointed interim co-lead counsel pursuant to Rule 23(g).

Rule 23(b)(3)’s predominance and superiority requirements. The court held that the plaintiffs satisfied Rule 23(b)(3)’s predominance and superiority requirements, emphasizing that antitrust cases are often well-suited for class treatment when liability, impact, and damages can be proven through common evidence. The central inquiry was whether the Publishers had advanced a classwide methodology to demonstrate antitrust impact and damages that was aligned correctly with their theory of unlawful tying. The court concluded that they had done so through expert testimony employing a benchmark methodology to measure overcharges allegedly caused by Google’s tying of its DFP ad server to the AdX exchange.

The court devoted extensive analysis to Google’s Daubert challenge, which argued that the expert’s comparison between open-web AdX transactions and certain mobile-app auction transactions was unreliable. Rejecting that argument, the court explained that benchmark analysis is a well-accepted method of proving antitrust damages and that the expert adequately justified his selection of a comparator market representing a “but-for” world free of the challenged tying restraints. The expert compared Google’s approximately 20% AdX take rate for open-web publishers to the 5–10% take rates Google charged in a segment of the mobile-app advertising market not subject to similar product ties, conservatively adopting a 10% benchmark supported by internal Google documents reflecting competitive pressure to align with industry rates once tying constraints were relaxed.

The court also considered Google’s objections that differences between mobile-app and open-web advertising, such as SDK usage, aggregated advertiser demand, analytics and security features, and competitive conditions, undermined the benchmark’s validity. It concluded that these arguments go to the weight of the evidence, not its admissibility, and do not defeat predominance. Because the expert acknowledged and accounted for technological and market differences, the court found no impermissible “apples-and-oranges” comparison. It determined that the analysis plausibly isolates the price effects of the alleged tying conduct, providing common proof of classwide overcharge.

The court also rejected Google’s contention that individualized inquiries would be required to determine injury or overpayment, noting that publishers lacked à la carte choices among AdX features and that any asserted value differences apply uniformly across the class. Disagreements over whether the alleged tying actually caused the higher take rate raise merits issues that can be resolved on a classwide basis. The court further found that the damages model is properly tailored to the Publishers’ tying theory and therefore satisfies Comcast Corp. v. Behrend, 569 U.S. 27 (2013), even if Google disputes the ultimate conclusions. Finally, the court held that a class action is the superior method of adjudication given the number of affected publishers, the complexity of the claims, and the efficiencies of resolving common liability and damages issues in a single proceeding. It thus concluded that Rule 23(b)(3) is satisfied for the AdX Publisher Class, subject to the exclusion of instream video ad sales.

MCMs. The court held that multi-customer management firms (MCMs) may be included in the AdX Publisher Class and that their participation does not defeat Rule 23’s adequacy or typicality requirements. Google contended that indemnification provisions in MCM agreements created a fundamental conflict, asserting that MCMs could be required to indemnify Google for damages awarded to the publishers they manage and therefore possess incentives adverse to the class. The court rejected this theory, explaining that Google’s interpretation would improperly require parties allegedly injured by Google’s antitrust violations to indemnify Google for its own unlawful conduct, a result likely unenforceable under California law and contrary to public policy. Because any injuries suffered by MCMs would be economically indistinguishable from those suffered by open-web publishers, and because the asserted indemnification conflict was legally implausible, the court concluded that MCMs’ interests are aligned with the class and that their inclusion in the certified AdX Publisher Class is appropriate.

AdSense Class denied. The court denied certification of the proposed AdSense Publisher Class, concluding that it failed to satisfy the requirements of both Rule 23(a) and Rule 23(b)(3). The court focused on the fact that The Progressive, the sole proposed class representative, was not situated like the putative AdSense publishers it sought to represent. Although The Progressive sold some advertising through an AdSense interface, approximately 99.99% of its ad sales were routed through Google’s DFP ad server, not through a stand-alone AdSense ad server. This reality undercut the core AdSense theory, that AdSense publishers were forced to transact exclusively through an AdSense server and auction platform, and placed The Progressive outside the heartland of the alleged misconduct.

Because The Progressive’s advertising practices materially differed from those of the proposed class, its claims did not arise from the same course of conduct. They would require defenses and factual inquiries unique to it. The court therefore held that The Progressive was neither a typical nor an adequate representative, and that these deficiencies alone required denial of class certification. As a result, the court rejected the AdSense Publisher Class in its entirety without reaching broader predominance or damages issues.

Advertiser Class rejected. Finally, the court denied the motion to certify the proposed Advertiser Class filed by plaintiff Hanson. The court first determined that the named plaintiff, Christopher Hanson, was not a typical or adequate class representative because his use of Google Ads was limited to a brief period in 2016, well before the implementation of key alleged anticompetitive practices, most notably Unified Pricing Rules in 2019. As a result, Hanson did not experience the market conditions central to the advertisers’ claims, rendering his alleged injuries and litigation incentives misaligned with those of most putative class members.

It reiterated that Hanson’s claims were premised on Google’s alleged auction-manipulation practices, Buy-Side Dynamic Revenue Share, Project Bernanke, and Unified Pricing Rules, but concluded that these theories cannot sustain class treatment. Project Bernanke, which originated in 2013 as an outgrowth of BSDRS, evolved through multiple iterations, including Global Bernanke, versions targeting advertisers using automated bidding tools, and later adaptations known as Alchemist and Bell v.2. Across these iterations, Project Bernanke allegedly increased the volume of transactions routed through AdX for Google Ads advertisers while consistently preserving Google’s 14% take rate.

A fundamental defect was temporal mismatch: Hanson exited the Google Ads market in 2016, before the adoption of Unified Pricing Rules and before later iterations of Project Bernanke that allegedly generated the most significant competitive harm, rendering his claims atypical and his interests misaligned with those of the proposed class.

The court further held that Hanson failed to show that antitrust impact and damages could be established through common proof. Although Project Bernanke overlapped in part with Hanson’s market participation, the record showed wide variation in advertiser outcomes across bidding strategies, pricing outcomes, and transaction pathways. Hanson’s expert sought to average these effects across advertisers, but the court found that this methodology obscured substantial individualized differences rather than resolving them, defeating Rule 23(b)(3)’s predominance requirement.

The court also emphasized that the proposed Advertiser Class was overinclusive. Project Bernanke applied only to AdX transactions, yet the class definition swept in advertisers who transacted on AdSense and third-party exchanges, many of whom were never exposed to the alleged misconduct. Compounding the problem, a majority of putative class members did not use Google Ads during key periods when Project Bernanke was in effect, undermining any inference of classwide injury.

Finally, the court identified additional individualized issues that would overwhelm common ones, including advertiser-specific exposure to the challenged practices and the need to determine whether individual advertisers agreed to arbitrate disputes with Google. Taken together, the lack of typicality and adequacy, the absence of common proof of impact and damages, the overbreadth of the class definition, and the prevalence of individualized issues compelled denial of Hanson’s motion for class certification in its entirety.

The Case is No. 1:21-cv-07034-PKC.

Judge: Castel, P.

Attorneys: Steven M. Jodlowski (DiCello Levitt LLP) for Sweepstakes Today, LLC. John E. Schmidtlein (Williams & Connolly LLP) for Google L.L.C.

Companies: Sweepstakes Today, LLC; Google L.L.C.

MainStory: TopStory Antitrust Advertising NewYorkNews GCNNews