Antitrust scrutiny of information exchanges and algorithmic pricing tools has intensified recently—from public enforcers and private plaintiffs alike—as companies across industries increasingly use algorithmic and AI-enabled systems to gather market data and generate pricing recommendations. These developments have generated significant uncertainty for businesses that rely on pricing and benchmarking tools, particularly given the still-unsettled state of the law.
For example, last year in the Gibson litigation (D. Nev.), the Ninth Circuit affirmed the district court’s dismissal, holding that plaintiffs failed to plausibly allege a horizontal agreement in an alleged algorithmic-pricing conspiracy case in the hotel industry.[1] By contrast, in RealPage (M.D. Tenn.), the court allowed a similar algorithmic-pricing claim to proceed in part—emphasizing that, unlike Gibson, the RealPage complaint “unequivocally” alleged a “melting pot” of confidential competitor data being used to generate rental rate recommendations for multifamily rental properties.[2]
Just last week, the Third Circuit joined the algorithmic pricing fray by reversing a district court’s dismissal of a hub-and-spoke theory, unanimously holding that plaintiffs plausibly alleged a price-fixing conspiracy in Cornish-Adebiyi v. Caesars Entertainment, Inc.[3]
In Cornish-Adebiyi, Plaintiffs, casino-hotel guests, alleged that defendants, Atlantic City casino-hotels and Cendyn (owner of the at-issue software, Rainmaker), continuously provided current, non-public pricing and occupancy data to Rainmaker to effectuate the alleged conspiracy.[4] The algorithm then processed the competitors’ data and recommended room rates that were uploaded into each casino-hotel’s booking and property management systems.[5] Plaintiffs alleged that the casino-hotels accepted Rainmaker’s recommended prices ~90% of the time and had to go through an “override” process to diverge from setting the recommended price.[6]
The Third Circuit’s opinion focused heavily on the software’s use of the real-time, non-public data shared by the defendant casino-hotels, coupled with plaintiffs’ other allegations that defendants knew or hoped that sharing such information would lead to “optimal,” higher prices. These allegations, the Third Circuit concluded, supported an inference of a common plan to set room rates based on the shared algorithmic recommendations.[7] The fact that each defendant retained ultimate pricing discretion, including the ability to reject and override the Rainmaker software’s recommendations, did not defeat plausibility at the pleading stage given other facts in the complaint. Specifically, the Third Circuit emphasized that compliance with the Rainmaker software’s recommendations was alleged to be as high as 90%, and that the software’s design made it difficult for the casino-hotels to use the specialized override process to reject the recommended price.[8] Similarly, the Third Circuit emphasized that prices are fixed when agreed upon (even if not followed without fail), and that the alleged high compliance rate, combined with plaintiffs’ other allegations about how Cendyn’s software works and how Cendyn acted as an alleged “hub,” supported an inference of broader agreement.[9]
While the Third Circuit’s opinion may seem directly at odds with the Ninth Circuit’s decision in Gibson, there is a key distinction between the opinions. While the Ninth Circuit in Gibson affirmed dismissal based on defendants’ licensing the same software, the court expressly noted that its analysis “might change” if plaintiffs had alleged that the software provider also shared competing hotels’ confidential information among the licensees—an allegation not clearly made in the Gibson complaint.[10] In Cornish-Adebiyi, by contrast, the Third Circuit found that plaintiffs plausibly alleged that such non-public, real-time confidential information was shared by the casino-hotels.In addition, the Third Circuit pointed to several other allegations demonstrating that the casino-hotel defendants and Cendyn knew or understood that using the software would allow defendants to set “optimal” (i.e., higher) prices. These distinctions—allegations that defendants shared non-public data through Rainmaker while knowing that they would collectively use and benefit from it—seemingly led to the different result.[11]
Although these cases involve the hospitality and multifamily industries, the underlying theory in both cases is not industry-specific: any business using tools that ingest competitively sensitive information (even via an intermediary) and produce recommended, dynamic prices in turn should assume that plaintiffs and enforcers may test hub-and-spoke theories—especially when the tool’s value proposition is marketed as leveraging non-public data submitted by competitors to recommend “better” or “optimal” pricing. The Third Circuit’s decision underscores that courts may be arriving at a new approach, at least on the pleadings, when assessing hub-and-spoke theories in the “context of AI-driven dynamic pricing” cases.[12] Specifically, the decision suggests that courts may be starting to consider such algorithm-pricing coordination theories plausibly alleged without sufficiently specific allegations of direct competitor-to-competitor communications—i.e., the “rim” of the spoke—when plaintiffs allege that the software functions as the exchange mechanism, and that the defendants who submitted their non-public, real-time information to the software knew or understood that the software’s pricing recommendations would allow for higher prices based on the use of such data.
Please reach out to the antitrust team at Sheppard if you have any questions about this opinion or how it might impact algorithmic pricing tools moving forward.
FOOTNOTES
[1] See Gibson v. Cendyn Grp., LLC, 148 F.4th 1069 (9th Cir. 2025), cert. denied, 224 L. Ed. 2d 502 (Apr. 20, 2026).
[2] In re RealPage, Inc., Rental Software Antitrust Litig. (No. II), 709 F. Supp. 3d 478, 512 (M.D. Tenn. 2023).
[3] See Cornish-Adebiyi v. Caesars Ent., Inc., No. 24-3006, 2026 WL 2182291, at *2 (3d Cir. July 29, 2026).
[4] Id. at *5.
[5] Id. at *6.
[6] Id.
[7] Id. at *11.
[8] Id.
[9] Id.
[10] Gibson, 148 F.4th at 1083.
[11] See Cornish-Adebiyi, 2026 WL 2182291, at *11.
[12] Id. at *10.