Cornish-Adebiyi: The Third Circuit Speaks on Pleading Standards for Challenges to AI-Driven Price Benchmarking

The Third Circuit has issued an important decision reversing the dismissal of a Sherman Act Section 1 complaint challenging the use of AI-driven pricing software for Atlantic City casino hotel rooms.

In Cornish-Adebiyi v. Caesars Entertainment, Inc., No. 24-3006 (3d Cir. July 29, 2026), the appellate court overturned the district court’s dismissal of a class action complaint alleging defendant casino hotel operators used defendant Cendyn Group’s revenue optimization software, Rainmaker, to fix and artificially inflate hotel room prices.

DISTRICT COURT’S DISMISSAL OF PLAINTIFFS’ CLAIMS

Plaintiffs alleged the following:

• Each casino hotel continuously uploaded real-time, nonpublic pricing and occupancy information into Rainmaker;
• Rainmaker combined each participant’s confidential information with similar data from competing casinos;
• The AI software generated “optimal” room prices multiple times per day based on those real-time, nonpublic inputs;
• Those prices were automatically uploaded back into each casino hotel’s reservation system;
• The casino hotels accepted Rainmaker’s recommendations approximately 90% of the time; and
• Rainmaker constrained the ability to deviate from its recommendations by requiring “special overrides” and tracking the frequency of the casino hotels’ use of such overrides.

The district court dismissed the complaint, reasoning that plaintiffs had failed to plausibly allege the “rim” necessary for a hub-and-spoke conspiracy—that is, an agreement among the competing casino hotels themselves (versus separate, unilateral decisions by each of them to use Rainmaker).

The district court reasoned that each had independently subscribed to Rainmaker at different points in time, each ultimately retained final pricing authority, and each was free to accept or reject Rainmaker’s suggested rates. As a result, the district court found insufficient “parallel conduct” between the defendant casino hotel operators from which any agreement between them to fix room prices could be inferred.

THIRD CIRCUIT REVERSAL

The Third Circuit disagreed, reversed the decision, and sent the case back to the trial court to continue case proceedings. The Third Circuit disagreed with the district court that the complaint did not allege parallel conduct by the casino hotel defendants and pointed to, inter alia, how each casino hotel allegedly:

• Used Rainmaker (irrespective of the date first subscribed) during the class period;
• Continuously supplied Rainmaker its real-time, confidential, nonpublic data on pricing and occupancy; and
• Effectively delegated pricing decisions to Rainmaker, accepting its recommendations roughly 90% of the time.

The Third Circuit held that the complaint adequately alleged parallel conduct and went on to find that the circumstantial evidence alleged in the complaint allowed a plausible inference that this parallel conduct was the result of collusion. The Third Circuit credited that among the circumstantial evidence alleged was:

• Action against economic self-interest. The court noted how casino hotel operators, during periods of low occupancy, would have extra economic incentive to reduce rates in order to fill rooms—especially since hotel guests’ gambling is a material source of revenue—but did not do so, a departure from their historical practices;
• Exchanges of real-time, nonpublic commercial information through Rainmaker;
• Opportunities to conspire at industry events;
• Presentations by Cendyn encouraging the casino hotels to avoid “price wars” and a “race to the bottom” by using Rainmaker; and
• Allegations that each casino hotel operator understood its competitors were also using Rainmaker to set prices based on each other’s real-time, nonpublic pricing and occupancy data.

Notably, the Third Circuit rejected defendants’ argument that their ability to retain final pricing authority defeated claims of reaching an agreement.

While the court noted that “an antitrust plaintiff’s allegations may well fail if the alleged colluders routinely varied from the recommended prices,” the allegations here were that 90% of the ultimate prices offered complied with those recommended by Rainmaker, coupled with further allegations that there were various mechanisms applied to make deviation from the recommendations difficult and that the level of deviation was monitored and scored.

COMPARISON TO THE NINTH CIRCUIT’S OPINION IN GIBSON V. CENDYN

The Ninth Circuit, in Gibson v. Cendyn Group, LLC, No. 24-3576 (9th Cir. Aug. 15, 2025), previously considered similar allegations challenging the use of Cendyn’s software in the Las Vegas market but upheld the district court’s dismissal of the complaint. Importantly, the plaintiffs in Gibson chose not to advance allegations that Cendyn and the Las Vegas casino hotel operators reached a hub-and-spoke conspiracy like the one before the Third Circuit.

As a result, in Gibson, the only question before the Ninth Circuit was instead whether the various vertical software licensing agreements between Cendyn and the casino hotels were themselves, in the aggregate, a violation of Section 1 of the Sherman Act.

The Ninth Circuit determined that they were not, as those agreements did not impose any requirement to charge any particular rate, to coordinate with any competitors, or to otherwise limit any competitive behavior, nor did the complaint contain any allegations (unlike the Cornish-Adebiyi complaint) that, alongside other measures, confidential pricing information was exchanged among the defendants through Cendyn’s software.

PRACTICAL TAKEAWAYS

As one of two recent federal appellate decisions directly addressing algorithmic pricing under Section 1 of the Sherman Act, the Third Circuit’s ruling is a significant development in this emerging area of law. There are a few key takeaways from this decision:

• Court decisions have focused particular attention on pricing recommendation algorithms that are plausibly alleged to use real-time, nonpublic, commercial information from multiple businesses as inputs. Consider the risks associated with any software tools that could meet that description.
• Companies should act independently and unilaterally in making their own ultimate pricing decisions. Consider adopting appropriate policies and procedures to align any use of algorithmic tools with this principle, including considering how best to appropriately document the unilateral basis for price decisions.
• The potential ability to reject the proposed price may not carry the day at the motion-to-dismiss stage. Where a proposed price is adopted the vast majority of the time, such as the 90% rate alleged for Rainmaker, and there are constraints on deviation, a complaint may survive, possibly leading to expensive, time-consuming antitrust discovery.
• Documentation of procompetitive benefits is crucial. Court rulings in algorithmic pricing antitrust cases have, to date, reached different conclusions on the appropriate standard for analyzing such claims. Courts applying the antitrust “rule of reason” framework will consider any potential procompetitive benefits of the challenged conduct if the case proceeds past the pleading stage. Ensure that any applicable procompetitive benefits of the algorithm, such as lowering prices for consumers or expanding output, are well documented.