U.S.

Concerns and challenges raised in the past few years by lawmakers, enforcers, and private plaintiffs about algorithmic pricing, particularly where algorithms are based on pooled competitor data, began to bear fruit in 2025 through significant court decisions, litigation activity, and legislation. These developments revealed fault lines between state and federal authorities, setting the stage for conflict between increasingly independent state competition regulation and the Trump administration’s aggressive press for a deregulatory national AI policy, discussed further in our AI Industry Preview. Notwithstanding this broader uncertainty, resolutions and decisions reached in certain real estate algorithmic pricing cases may provide a basis for enforcers and plaintiffs to extend their logic to algorithms that influence competitive decision-making in other domains.

Legal scrutiny of algorithmic pricing and shared-data algorithms was kickstarted by a 2022 report on RealPage’s rental price management software and has continued to center on both RealPage specifically and the real estate sector more generally. Classes of private plaintiffs fired the first shots, and their suits were consolidated into multidistrict legislation in the Middle District of Tennessee in September 2023.[1] State attorneys general moved first among enforcers, with the District of Columbia filing in November 2023 and Arizona filing in February 2024—both in state court. Both suits remain active, without any major rulings having been issued.

The U.S. Department of Justice (DOJ) filed its own suit under the Biden administration in August 2024, joined by nine state attorneys general. The DOJ shortly thereafter closed a parallel criminal investigation into RealPage without taking any action. The DOJ complaint alleged that RealPage pooled nonpublic competitively sensitive transactional data from its customers to create recommendation algorithms, including an AI model, and induced customers to accept pricing recommendations. The complaint notes RealPage statements that its tools were intended to support higher pricing across the industry as a whole.

Given the Trump administration’s strong stance on promoting the development and use of AI technologies by American businesses, there was some question as to whether the DOJ’s case against RealPage would continue. The Trump administration’s DOJ did not withdraw the case, but settlements with Greystar and Cortland in separate cases alleging collusion via RealPage software suggested it may have been seeking to settle with RealPage as well.

Indeed, a proposed settlement was announced in November 2025. The settlement, in broad terms, limits RealPage’s ability to collect certain kinds of data, with different limits set for AI model training versus runtime operation, restricts the granularity of recommendations RealPage can provide, and prevents RealPage from inducing recommendation compliance or suppressing recommendations to lower price. The structure and limits of the settlement provide guidance on how federal enforcers will evaluate shared-data algorithms and AI models going forward. However, the state plaintiffs did not join the proposed settlement and have not yet announced a course of action.

Since RealPage, several cases have been filed challenging the use of pricing algorithms (whether they rely on pooled confidential data or not) used to set or influence pricing decisions, both in real estate or rental pricing and in other industries such as healthcare reimbursement rates.[2] Courts have reached varying outcomes on how to evaluate such algorithms as a matter of law.

For instance, in the private RealPage litigation, the plaintiffs alleged both a per se and a rule of reason theory. Although the Biden administration’s DOJ filed a statement of interest in favor of per se treatment, the U.S. District Court for the Middle District of Tennessee dismissed that count and allowed only a rule of reason theory to proceed, citing a lack of judicial experience with pricing algorithms. The U.S. District Court for the Western District of Washington expressly disagreed with that reasoning and held in Duffy v. Yardi that mere use of a novel technology to effectuate a horizontal price fixing conspiracy did not remove it from per se treatment.[3]

Courts have also reached different outcomes on the sufficiency of allegations of an agreement required to state a claim under Section 1 of the Sherman Act. The Yardi court denied a motion to dismiss, highlighting allegations that the lessor defendants understood that Yardi would use commercially sensitive information they provided in order to recommend rates that would maximize rents across the industry. The court buttressed its conclusion by pointing to plus factors such as Yardi marketing materials and internal lessor documents recognizing the price increasing impact of the tool. However, the U.S. District Court for the Northern District of California in Dai v. SAS Institute granted a motion to dismiss, holding insufficient allegations that the relevant software merely used data from its customers without providing details on whether and how such data was used to produce the ultimate recommendations. These cases suggest a high bar for plaintiffs, spurring responsive legislation as discussed below.

The U.S. Court of Appeals for the Ninth Circuit decision in Gibson v. Cendyn Group in August 2025 was the first (and remains the only) appellate decision on the topic of pricing algorithms. The plaintiffs alleged that the hotel defendants had unlawfully conspired to raise room prices through use of a common price recommendation service. The Ninth Circuit upheld the district court’s dismissal and imposed a limited bar on stating claims based on algorithmic pricing tools, holding that allegations that the defendants knowingly subscribed to the same pricing recommendation tool and that prices rose after those subscriptions were not enough to state a claim.

The decision leaves open some important questions. The court itself acknowledged that the analysis may be different where i) the defendants agree to take or rely on tool recommendations or where ii) the algorithm pools, share, or uses data provided by one subscriber to make recommendations for another. In addition, the Biden administration’s DOJ had filed an amicus brief in October 2024 in support of the plaintiffs, arguing that the district court had erred by seeming to require proof that an agreement eliminated all pricing discretion to be unlawful—a question not reached by the Ninth Circuit decision but that has remained significant in legislation and DOJ settlements, even in the Trump administration.

Legislative efforts to address real estate pricing algorithms started to gain traction shortly after government enforcement ramped up. Starting with San Francisco in October 2024, a number of municipalities adopted bans on algorithmic rent-setting tools. Certain state legislatures have since followed suit, with California codifying the application of its antitrust laws to shared-data algorithms in general and New York passing an outright bar on such algorithms in real estate markets—both in October 2025. The Colorado governor vetoed a bill similar to New York’s in May 2025.

Both RealPage and the Trump administration have pushed back on this legislation. RealPage filed suit to challenge a Berkeley ordinance that would have prohibited the use of RealPage in April 2025, arguing inter alia that the law amounted to an unconstitutional restriction on RealPage’s first amendment rights. In response, Berkeley voted to delay implementation of the ordinance until 2026 to allow other cases involving RealPage to resolve. In November 2025, RealPage lodged a similar challenge to New York’s state-level law.

The Trump administration has sought to establish a national AI policy that would preempt state legislation (see also our AI Industry Preview). Although the administration failed to include a state AI regulation moratorium in the “One Big Beautiful Bill,” the idea resurfaced in another form later in the year. In December 2025, the administration issued an Executive Order (EO), “Ensuring a National Policy Framework for Artificial Intelligence.” Among other things, the EO directs federal authorities to identify and challenge state laws conflicting with national AI policy as unconstitutional or preempted by federal law or regulation. The scope of this instruction is not yet certain, but it could extend to regulations affecting the use of pricing algorithms based on AI.

We expect courts, agencies, and lawmakers to continue to grapple in the coming year with some of the points of tension that have been revealed in this year’s activity, including, but not limited to:

• the extent to which increasingly sophisticated AI models can provide actionable recommendations based on categories of data previously deemed lower risk, such as public, historical, or aggregated data;
• whether and how legal distinctions should be drawn between sharing and using data for AI model training versus during algorithm runtime;
• how much independent review of or discretion in adopting shared-data algorithm recommendations is sufficient to avoid an inference of unlawful agreement; and
• whether federal AI policy can preempt restrictions on shared-data algorithms that use AI techniques.

We anticipate that federal enforcers will continue to assess these questions through the lens of information-sharing precedents and expect an emphasis on reaching settlements modeled on RealPage that seek to delineate between permissible and anticompetitive exchange. By contrast, we expect the state legislatures and enforcers that have led the charge thus far to defend their prerogatives in the face of federal pressure and to deploy new tools and theories beyond traditional Section 1 information-sharing analysis. Judicial guidance in this area is likely to lag enforcement, particularly given the settlement of the DOJ’s RealPage case. This gap, combined with dueling policies at the state and federal level, creates risk of uncertain or even conflicting requirements on the developers and users of shared-data algorithms, requiring careful counsel from experienced practitioners.

Europe and the UK

In Europe, investigations of algorithmic pricing are still in their infancy. While multiple competition authorities have expressed great interest in algorithmic pricing, investigations are still comparatively rare and have not yet resulted in final decisions.

In July 2025, a senior official of the European Commission (EC) confirmed at a conference that the EC was looking at multiple algorithmic pricing antitrust investigations. In September 2025, the president of Poland’s antitrust authority (UOKiK) confirmed it was investigating instances of potential collusion via algorithmic pricing tools in the banking and pharmaceutical sectors. UOKiK’s president stated in an interview that banks may be using algorithms fed by data from Poland’s largest credit risk database as well as from internal sources to coordinate pricing in consumer loans and mortgage markets. In the pharmaceuticals sector, UOKiK suspects three major wholesalers (controlling 80 percent of the market) may be using IT systems to exchange commercially sensitive information on drug prices, margins, and volumes sold in affiliated pharmacies.

The German Federal Cartel Office’s (FCO) investigation into Amazon’s algorithmic price control mechanisms is more advanced, and there somewhat more publicly available information as to its scope and status. In June 2025, the FCO announced that it had sent formal charges to Amazon regarding its price control mechanisms for third-party sellers on its platform. The FCO alleged that Amazon relies on algorithms to set dynamic price caps and benchmark third-party offers against them. The FCO claims that offers considered too expensive may be delisted or otherwise disadvantaged (including loss of the Buy Box, lower search ranking, or advertising ineligibility), and traders may be prompted to align their prices with Amazon’s benchmarks.

The Netherlands Authority for Consumers and Markets (ACM) has launched a market investigation into algorithmic pricing practices in the airline industry, which has been long associated with personalized and dynamic pricing. Only one day later, Italy’s competition authority, the AGCM, also indicated that it was engaging the EC over how to improve the price comparison of airline tickets. This follows the AGCM’s own market investigation into how airlines employ algorithmic pricing, the preliminary findings of which were published in 2024.

Similarly, while the UK’s Competition and Markets Authority (CMA) has professed an intention or expectation to consider algorithmic pricing cases, it has not yet launched any investigations, and it remains to be seen if/when the CMA takes action in this area.

European and UK antitrust authorities are clearly poised to take steps on algorithmic pricing in the future, but at present cases are at an earlier stage than those in the U.S. It is likely that 2026 will be a key year for the development of algorithmic pricing cases in these jurisdictions.

For more information on these developments or to discuss strategies for limiting the risks posed by shared-data algorithms, please contact any member of the firm’s Antitrust and Competition practice.

---

[1] Wilson Sonsini currently represents a defendant in this litigation and in two related matters.

[2] The case, In re MultiPlan Health Insurance Provider Litigation, survived a motion to dismiss in June 2025. The Trump administration’s DOJ filed a statement of interest to assert its view that a violation may be found even where data is shared through an intermediary and even where users do not necessarily use a common algorithm in the same way, mirroring the Biden administration’s DOJ’s amicus brief in Cendyn, discussed below.

[3] The Biden administration’s DOJ filed a statement of interest in this case on the same grounds as in RealPage.