Recent developments in algorithmic pricing: U.S. appeals court weighs in, enforcers stay aggressive, and open questions remain

The first federal appellate decision on algorithmic pricing provides clarity on some, but not all, of the many challenging legal issues surrounding these practices

In Gibson v. Cendyn Group, the Ninth Circuit¹ agreed with a lower court ruling² that it does not violate Section 1 of the Sherman Act for competing hotels to independently use the same third-party pricing software to generate recommended pricing for their hotel rooms, at least where there was no underlying agreement amongst the competitors to do so, and where the software did not share any licensing hotel’s confidential information among competing licensees. The Ninth Circuit found that not even a rule of reason analysis is warranted in cases where there is no causal link between the purported agreement and an anticompetitive restraint. The Gibson decision marks the first time a U.S. appellate court has applied the antitrust laws to cases involving algorithmic pricing systems and is a significant addition to the rapidly developing caselaw addressing these issues.

At the initial stage of the case, the District of Nevada ruled that the plaintiffs had failed to plead that the hotel defendants’ independent licensing agreements to use a common pricing software was equivalent to an agreement on price.³ On appeal, the Ninth Circuit ruled that the independent licensing agreements with software provider Cendyn do not constitute a restraint of trade because: (1) the defendants did not agree amongst each other to follow Cendyn’s pricing recommendations (and parallel conduct absent an agreement is insufficient to state a claim under Section 1); and (2) “antitrust law provides no mechanism by which courts can evaluate the independent agreements between Cendyn and each Hotel Defendant ‘in the aggregate.’”⁴ The Ninth Circuit rejected the plaintiffs’ characterization of the contracts as vertical agreements between Cendyn and the hotels, since “relative to Hotel Defendants, Cendyn is not up or down the supply chain in the relevant market: hotel-room rentals on the Las Vegas Strip . . . [and] the software is not an input that goes into the production of hotel rooms for rentals.”⁵ The Ninth Circuit also concluded that, while “increased prices can serve as evidence of ‘a substantial anticompetitive effect’ in the context of certain agreements . . . that does not mean an individual firm’s independent choice to charge higher prices itself harms competition . . . [and] charging a higher price by itself is not anticompetitive.”⁶

The Ninth Circuit’s ruling provides some clarity on whether and how the use of algorithmic pricing software may violate the antitrust laws, and what the relevant considerations will be for courts making this determination. However, as the court pointed out, the plaintiffs had not “allege[d] that Cendyn pools, shares, or uses the confidential information provided by a given Hotel Defendant into the pricing recommendations it generates for any other Hotel Defendant.”⁷ Accordingly, the Ninth Circuit’s opinion did not address the antitrust implications of a tool that generates pricing suggestions based on the confidential information from competitors. In addition, the Ninth Circuit was not tasked with reviewing the district court’s dismissal of the plaintiffs’ allegation that the hotels entered into a “hub-and-spoke” agreement to utilize Cendyn’s revenue-management software and adopt the algorithm’s pricing recommendations.⁸ But the court made clear that “it would undoubtedly violate Section 1” if the hotels “agree among themselves to abide by a third party’s pricing recommendations,”⁹ and that such an agreement “would be a hub-and-spoke conspiracy”¹⁰ that would be “analyzed according to the per se rule and the rule of reason, as applicable.”¹¹

A very similar case is now on appeal to the Third Circuit, after the District of New Jersey granted a motion to dismiss in a lawsuit targeting different hotels’ use of the same software at issue in Gibson. Like the lower court in Gibson, the New Jersey court ruled that, because the plaintiffs did not allege that the algorithmic software at issue used competitors’ nonpublic, proprietary data to generate pricing recommendations, they failed to allege an illegal hub-and-spoke conspiracy.¹²

Lower courts continue to debate whether to apply the per se approach or rule of reason

The question left open in Gibson—whether to apply the per se rule or the rule of reason in cases alleging price fixing through algorithmic pricing—continues to vex the lower federal courts.

In April 2023, a federal court in Tennessee ruled in the ongoing RealPage litigation—a private plaintiff class action challenging multifamily residential lessors’ use of RealPage’s revenue management software—that it is not per se illegal for competitors to use a common pricing tool, at least where there was no direct evidence of a direct agreement among the competitors, and where plaintiffs failed to allege that RealPage or any of the lessors “can enforce acceptance of price recommendations” provided by the algorithmic pricing tool.¹³ Instead, the judge held that a rule of reason analysis applied to plaintiffs’ Section 1 claim, and in applying that standard, denied defendants’ motion to dismiss.¹⁴ Then, a few months later in Duffy v. Yardi Systems, a case involving similar allegations, a federal judge in Washington held that where plaintiffs adequately alleged “both invitation and acceptance and sufficient plus factors to give rise to a plausible inference of a preceding agreement” and that defendants implemented the software’s pricing recommendations, the per se standard should apply.¹⁵ Judge Lasnik noted in his Duffy opinion that he “respectfully disagrees” with the RealPage ruling, saying that “[t]he point of the per se rule and rule of reason analysis is not to create a heightened pleading standard related to the existence or purpose of the conspiracy, but to determine whether the restraint on trade is ‘unreasonable’ for purposes of § 1. Chief Judge Crenshaw [in RealPage] cites no authority for judging the reasonableness of an adequately alleged conspiracy to restrain trade by the strength of the conspiracy allegations.”¹⁶

More recently, in June 2025, a federal court in Illinois denied a motion to dismiss a consolidated class action complaint alleging that defendants—health insurance data analytics firm MultiPlan and several health insurance companies—unlawfully conspired to underpay out-of-network health care providers by outsourcing rate-setting to MultiPlan.¹⁷ The court found that that plaintiffs had sufficiently alleged that third-party payors who contracted to use MultiPlan’s rate calculation software and negotiation services collectively agreed to fix prices for out-of-network service payments by sharing sensitive information that MultiPlan used in its algorithmic software to “align” rates for payors. The court in its ruling applied the per se standard, noting that “horizontal price-fixing agreements are per se illegal,” and that “the plaintiffs have plausibly alleged a horizontal hub-and-spokes price-fixing agreement.”¹⁸

DOJ and State AGs continue to pursue antitrust enforcement targeting algorithmic pricing tools, and early settlements may shed light on evolving standards

Under the leadership of AAG Abigail Slater, DOJ has reasserted a commitment to antitrust enforcement targeting “algorithmic collusion” that was first established during the Biden administration. In a recent video statement on social media, AAG Slater said that she expects the number of DOJ Antitrust Division investigations into potential collusion through algorithmic pricing to grow, and that DOJ is “not afraid to litigate” these cases. Slater also advised firms to “perform their own due diligence on shared algorithms inputs and functionality to prevent collusion that can harm consumers.”¹⁹

DOJ’s lawsuit against RealPage and six multifamily residential lessors is currently making its way through federal court in North Carolina. The government alleges that RealPage violated both Section 1 and Section 2 of the Sherman Act because: (1) RealPage’s revenue management software collects competitively sensitive data from landlords and generates rental pricing recommendations; and (2) the landlords agreed both with RealPage and each another to share nonpublic, sensitive information, both indirectly through the RealPage software, and directly.²⁰ The amended complaint also alleges that, in combination with the use of RealPage’s pricing algorithms, the defendants’ coordination is “even more likely to restrain, rather than promote, competition.”²¹ But as we outlined in a prior client alert, DOJ’s amended complaint against RealPage—which ten states have joined as co-plaintiffs—does not allege that the conduct is a per se violation of Section 1.²² In other recent comments, AAG Slater noted that “there’s nothing wrong with using algorithms,” but that the RealPage tool “was running on proprietary, competitive, sensitive data that went to output [and] price,” and that where certain “parameters are in place”—like the algorithm’s being shared by competitors and “relying on sensitive data”—then “it can lead to bad, anticompetitive outcomes.”²³

On August 8, 2025, DOJ announced²⁴ a settlement with one of the lessor defendants, which may provide some guidance on what DOJ believes is an appropriate use of pricing algorithms. If approved, the proposed final judgment²⁵ would, among other provisions, prevent use of any revenue management product that uses external, nonpublic data to generate rental prices or rental pricing recommendations, or uses a model or algorithm trained on nonpublic data.²⁶

Nevada is not a co-plaintiff in the DOJ’s case against RealPage, but the Nevada attorney general recently announced that it had entered into a consent judgment with RealPage in state court, agreeing to settle allegations of anticompetitive conduct involving use of the software in the state. The settlement is RealPage’s first in this wave of antitrust enforcement and litigation, and while RealPage provided no admission of wrongdoing or liability, the settlement terms limit RealPage to only:

• Use the nonpublic data of other unaffiliated properties in the runtime operation of its revenue management software to calculate recommendations for properties in Nevada if the data is at least three months old, anonymized, and aggregated to at least ten properties;
• Publish nonpublic data regarding rent, occupancy or availability to its customers in Nevada if the data is aggregated to at least ten properties and anonymized;
• Use machine learning models trained with nonpublic data from properties in Nevada if the data is aged at least three months.²⁷

States and municipalities seek legislation banning residential real estate companies’ use of algorithmic pricing tools

Over the past year, many states and municipalities have introduced ordinances prohibiting the use of certain algorithmic devices or services to set rents or occupancy levels for residential rental units. The San Francisco Board of Supervisors enacted one of the first such ordinances in October 2024.²⁸ Similar ordinances are currently in effect in Providence, Rhode Island; Hoboken, New Jersey; Jersey City, New Jersey; Seattle, Washington; San Diego, California; Minneapolis, Minnesota and Philadelphia, Pennsylvania.²⁹ Generally speaking, these ordinances prohibit the use or provision of algorithmic tools that incorporate certain nonpublic competitor information provided by competing real estate lessors to generate suggested rental prices, lease terms, and/or occupancy levels for users.³⁰ What qualifies as “nonpublic competitor information” under these ordinances varies by municipality. For example, Jersey City’s ordinance limits its ban to those algorithmic pricing systems that produce rental pricing or occupancy recommendations based on nonpublic competitor information that is less than 365 days old;³¹ the statutory definitions of “nonpublic competitor information” in the San Diego³² and Minneapolis ordinances³³ do not include that sort of date limiter.

Similar laws have been introduced in the state legislatures in California, Illinois, Massachusetts, Michigan, New Jersey, New York, North Carolina, Ohio, Pennsylvania, and Wisconsin. To date however, no state has enacted a law banning the use or sale of algorithmic pricing devices in residential real estate markets. The Colorado state legislature passed a law that would have prohibited the sale or use of an algorithmic device to competing landlords that recommends and sets the amount of rent, level of occupancy, or other commercial term associated with the occupancy of a residential premises; however, Governor Jared Polis vetoed the bill in May 2025.³⁴

Takeaways

With the state of play continuing to evolve, few certain predictions can be made about what is next in antitrust enforcement against algorithmic pricing tools. But there are at least some clear takeaways from these recent developments:

• First, although labels like “algorithmic pricing” and “A.I. tool” are used often and interchangeably, these technologies do not all offer the same functionality or present the same antitrust risk. Companies relying on pricing tools that might fit under the “algorithmic pricing” umbrella must have an exacting understanding of those tools' terms of use, including, e.g., whether the tool automatically implements pricing recommendations, and the sources of the information provided to the tool.
• Second, early indications that courts might be hesitant to apply the per se standard in any cases involving algorithmic pricing, because the technology and practice was relatively new, have not endured, and several federal courts have allowed per se claims to proceed past motions to dismiss.
• Third, despite the modern technology at issue in these cases, courts seem to be looking for familiar hallmarks of potentially anticompetitive conduct—agreements among competitors to, e.g., follow the recommendations of a shared tool, and an exchange of competitively sensitive information. Even where courts do not find the per se standard appropriate, the nature of a given market (i.e., the degree of concentration, barriers to entry) can elevate the antitrust risk.
• Finally, although the federal agencies usually lead on antitrust enforcement priorities, amidst some relative uncertainty in the caselaw, state and local authorities are increasingly looking to ramp up enforcement, even if that means pushing the law to more aggressively police these practices. It is essential that companies who may have antitrust risk exposure track not only DOJ and FTC developments, but also be mindful of nuanced distinctions in new state and local laws as well.

For companies that are currently using or considering investing in algorithmic pricing tools, experienced antitrust counsel like the team at Hogan Lovells can help navigate the enforcement and legal uncertainty and help you understand how developments at the federal, state, and local level might impact your business's risk exposure.

Authored by Ben Holt, Claude Szyfer, Holden Steinhauer, and Jill Ottenberg.