Algorithmic Collusion Hits a Circuit Split: Third Circuit Revives AI Price-Fixing Class Action

The End of the Smoke-Filled Room

For decades, antitrust litigators hunting for Section 1 Sherman Act violations have looked for the proverbial smoke-filled room—emails, recorded calls, or secret meetings where competitors agreed to fix prices. But what happens when the cartel is just a line of code? On July 29, 2026, the U.S. Court of Appeals for the Third Circuit dragged antitrust law fully into the algorithmic age, reviving a proposed class action alleging that casino operators used artificial intelligence revenue-management software to coordinate room pricing and overcharge guests.

This decision is a seismic event for the plaintiffs’ bar and corporate defense counsel alike. Why? Because it formally opens a circuit split on the hottest issue in antitrust law: algorithmic price-fixing. As Reuters reported, the Third Circuit’s revival of the New Jersey casino litigation directly diverges from a recent Ninth Circuit decision that tossed a nearly identical class action against Nevada casinos. The courts of appeals are now fundamentally divided on what constitutes a "conspiracy" when competitors outsource their pricing autonomy to a shared, third-party AI platform.

Delegation as Collusion: The Hub-and-Spoke Dilemma

Under Section 1 of the Sherman Act, 15 U.S.C. § 1, plaintiffs must prove a "contract, combination... or conspiracy, in restraint of trade." In the context of motions to dismiss, Bell Atl. Corp. v. Twombly, 550 U.S. 544 (2007), requires plaintiffs to plead enough factual matter to suggest that an agreement was made, rising above mere parallel conduct.

Algorithmic pricing cases rely on a "hub-and-spoke" conspiracy theory, a doctrine tracing back to Interstate Circuit, Inc. v. United States, 306 U.S. 208 (1939). The software vendor is the "hub," and the competitor casinos are the "spokes." But a hub-and-spoke conspiracy fails without a "rim"—a horizontal agreement among the competitors themselves. The Ninth Circuit looked at AI pricing in Nevada and found no rim, ruling that casinos independently choosing to buy the same commercially available software does not constitute a horizontal conspiracy.

The Third Circuit, however, has taken a decidedly more plaintiff-friendly posture. By reviving the New Jersey casino suit, the court suggests that when direct competitors knowingly feed their proprietary data into a shared algorithm that dictates market-wide pricing, the inference of a horizontal agreement is plausible enough to survive a Rule 12(b)(6) challenge.

For practicing lawyers, the takeaway is absolute: Using third-party revenue-management software is no longer a safe harbor from antitrust liability. If your client uses algorithmic pricing models that aggregate data from horizontal competitors, they are operating in the danger zone.

Class Certification: The Coming Daubert Bloodbath

While the Third Circuit’s decision makes it easier for plaintiffs to unlock discovery, the true war in these AI antitrust cases will be fought at the class certification stage under Fed. R. Civ. P. 23(b)(3). And another recent appellate development signals just how brutal that fight will be.

On August 24, 2026, the Seventh Circuit agreed to hear an immediate Rule 23(f) appeal from Cornell and other elite universities challenging a class certification order in a massive financial-aid antitrust lawsuit. Reuters noted that the Seventh Circuit specifically limited its review to whether the district court adequately analyzed expert testimony on the common proof of antitrust impact.

This Seventh Circuit development is intimately connected to the future of the Third Circuit's casino litigation. In any algorithmic price-fixing case, plaintiffs cannot rely on traditional anecdotal evidence of harm; they must use complex econometric models and dueling AI experts to prove that the algorithm artificially inflated prices for the entire class. Under Comcast Corp. v. Behrend, 569 U.S. 27 (2013), a model for determining class-wide damages must precisely match the plaintiff's theory of liability.

The Seventh Circuit’s willingness to scrutinize the district court's gatekeeping role over expert testimony at the class-cert stage is a warning shot. It tells defense counsel that even if you lose the motion to dismiss in an algorithmic collusion case, you can still defeat the class by aggressively challenging the plaintiffs' econometric models under Daubert during the Rule 23 fight.

Looking Ahead: The Supreme Court Must Intervene

The Supreme Court has historically been picky about antitrust class actions. On April 20, 2026, the Court declined to review a bank-collusion class-action dispute, leaving intact class-action treatment for a $12 billion case brought by cities against major banks. But algorithmic price-fixing is different. It is a novel, structural shift in the American economy.

With the Third and Ninth Circuits now split on whether the mere use of shared pricing algorithms constitutes a Sherman Act conspiracy, a Supreme Court showdown is inevitable. Until then, corporate counsel must audit their clients' use of AI pricing tools, specifically looking for "data pooling" features where the algorithm trains on competitors' non-public pricing data. Litigators, meanwhile, should prepare for a massive influx of Section 1 class actions filed in the friendly confines of the Third Circuit.