Leaving Las Vegas: Third Circuit Breaks From Ninth Circuit Ruling on Algorithmic Pricing Claims

On July 29, 2026, in Cornish-Adebiyi v. Caesars Entertainment, Inc., the Third Circuit reversed the dismissal of a putative class action alleging that Atlantic City casino-hotels conspired to fix guest-room rates through a common pricing algorithm.¹ It is the first federal appellate court to allow algorithmic price-fixing claims to survive a motion to dismiss. The decision comes less than a year after the Ninth Circuit became the first federal appellate court to address the antitrust implications of algorithmic pricing in Gibson v. Cendyn Group, LLC — where it upheld the district court’s dismissal of a complaint involving Las Vegas casino-hotels using the same algorithmic pricing software.² While both cases involved the use of the same software, the differing outcomes illustrate not a conflicting approach to the applicable legal standards but to differences in the facts alleged. The Third Circuit found a viable antitrust claim because it concluded that the complaint adequately alleged a horizontal agreement among competitors and that the software used competitors’ non-public, competitively sensitive information to recommend pricing.

Background

In Cornish-Adebiyi, Plaintiffs allege that each casino-hotel fed its current, non-public pricing and occupancy data into the pricing software, which processed that data alongside competitors’ data to generate recommended room rates that the hotels accepted roughly 90% of the time, producing rates that rose in parallel even as occupancy fell.³ In September 2024, the District of New Jersey dismissed the complaint, following the district court’s reasoning in Gibson and finding no adequate allegation of a “rim” to the alleged hub-and-spoke agreement (i.e., no horizontal agreement among the hotels).⁴ The court cited the hotels’ staggered adoption of the software over a 14-year period, the absence of allegations that data was “pooled” or “commingled,” and the hotels’ retention of final pricing authority.⁵

Third Circuit Opinion

The Third Circuit reversed the dismissal, holding that the allegations, taken as true, plausibly support an inference that the competitor hotels agreed to fix room rates through the software.⁶ The opinion reasoned that AI-enabled pricing tools could bridge the communication and monitoring gaps that historically made collusion difficult, allowing competitors to coordinate prices “without ever communicating with each other” and to police deviations in real time.⁷ At the same time, the court took pains to note that there is nothing inherently unlawful about using pricing algorithms, and that its decision made no findings about how the software actually works.⁸

According to the court, the complaint adequately alleged that each hotel understood that its pricing recommendations were built on the non-public data all users supplied and that each “was committed to a common plan of setting room rates based on the recommended rates received from [the] software … while also ‘knowing that their competitors would not lower their room rates to take market share.’”⁹ While the court acknowledged that information exchanges even among competitors are not per se unlawful, it held that the alleged exchanges of current pricing and occupancy data were adequately alleged to “faciliat[e] collusive conduct.”¹⁰ The court declined to require at the pleading stage allegations explaining how the algorithm worked to facilitate the alleged collusion. Nor did the hotels’ ability to override recommendations defeat the claim because the alleged friction of “requiring a special override to be used only in times of ‘need and extreme circumstances,’ and [] scoring each casino-hotel on how often it overrides the algorithm’s price and forecast recommendation” combined with the alleged approximately 90% adherence rate to recommended prices supported an inference of agreement.¹¹

In conclusion, the Third Circuit distinguished the pricing algorithm from using “identical spreadsheets or pricing formulas” or even using software to set production levels because it allegedly involved “exchange of non-public commercial information” that was then used to provide price recommendations that were adopted most of the time.¹² While the court acknowledged the potential pitfalls inherent in assessing the competitive impacts of new and shifting technologies, it ultimately concluded that “the goal of our antitrust jurisprudence remains the same: to ensure the continued existence of ‘independent centers of decision-making,’” and held that “[the software] is alleged to operate as a single decision-maker or hub, coordinating pricing for a majority of the market.”¹³

How Cornish-Adebiyi Differs From Gibson

Although both Cornish-Adebiyi and Gibson reached different results analyzing the same software, the two appellate decisions are largely reconcilable — and the differences are instructive.

First, the theories on appeal diverged. In Gibson, the plaintiffs abandoned their “hub-and-spoke” theory and pressed only the argument that the hotels’ separate license agreements with a common vendor, in the aggregate, violated the antitrust laws. The Ninth Circuit rejected that theory, describing the license agreements as ordinary sales contracts that did not restrain any hotel’s ability to price independently. In Cornish-Adebiyi, by contrast, the plaintiffs pursued the horizontal (hub-and-spoke) agreement theory and supported it with alleged parallel conduct and plus factors.

Second, the allegations differed as to the sharing of non-public, competitively sensitive information. The Gibson software was not alleged to pool, share, or otherwise use one hotel’s confidential information to inform prices for competitors, and the Ninth Circuit expressly signaled that its analysis may be different if the software allowed the sharing of confidential information among competitors.¹⁴ In Cornish-Adebiyi, the Third Circuit found that the complaint alleges that the algorithm ingested each hotel’s non-public, real-time pricing and occupancy data and used that pooled data to inform price recommendations to competitors.

Third, the Cornish-Adebiyi complaint alleged facts sufficient to infer agreement because the hotels accepted the algorithm’s recommended rates roughly 90% of the time and could override them only in times of “need and extreme circumstances.”¹⁵ By comparison, Gibson found that the complaint in that case did not allege that the hotels were required to accept the software’s recommendations.    

In other words, rather than taking a conflicting approach, the Third Circuit relied on allegations that the Ninth Circuit found relevant but absent in Gibson. The two decisions together confirm that allegations concerning (a) a horizontal agreement among competitors (a “rim”), (b) an exchange of non-public, competitively sensitive information, and (c) users’ adherence to the software’s recommendations may be the difference between dismissal and proceeding to discovery.

Takeaways

Cornish-Adebiyi reinforces the guideposts our August 2025 Advisory drew from Gibson and shows how these key issues can be outcome determinative. Companies evaluating algorithmic pricing tools should continue to ask:

• Was the decision to use the software reached independently? Communications or understandings with competitors about whether to use a tool, which tool to use, or how to use its outputs can supply the “rim” to an alleged antitrust conspiracy.
• Does the tool incorporate non-public competitor data to make pricing recommendations? The Third Circuit relied heavily on the allegation that the software pooled each hotel’s confidential, real-time, pricing or occupancy data, and used that data to inform pricing recommendations to competitors. 
• Does the software or the license agreement limit the user’s ability to set pricing? The Third Circuit ruled that a high adherence rate and friction around overriding recommendations supported an inference of agreement, notwithstanding that each hotel retained nominal final pricing authority. 
• Does the software or license restrict the user’s ability to compete in some way? Terms that limit independent pricing or the ability to use other vendors suggest potential for anticompetitive effects.

As we cautioned after Gibson, enforcers have urged courts to treat the common use of pricing algorithms as potentially unlawful, and different district courts continue to reach different results.¹⁶ Cornish-Adebiyi now provides appellate support for plaintiffs where the allegations include contemporaneous use of software among competitors, the exchange of non-public competitively sensitive information, and high adherence to the software’s recommendations. The decision resolves the case only at the pleading stage, however, and the plaintiffs will face a higher burden to support their claims with evidence as the case proceeds. Nevertheless, companies should treat the opinion as a clear signal that third-party pricing tools can create antitrust risk and engage antitrust counsel to evaluate pricing software before adoption.

© Arnold & Porter Kaye Scholer LLP 2026 All Rights Reserved. This Advisory is intended to be a general summary of the law and does not constitute legal advice. You should consult with counsel to determine applicable legal requirements in a specific fact situation.