A class action lawsuit filed in California on June 22, 2026 accuses BP, Marathon Petroleum, 7-Eleven, Walmart, Speedway, and Albertsons of using artificial intelligence to coordinate gasoline and diesel prices in violation of federal and state antitrust law, according to Reuters.

The complaint alleges that the AI-facilitated scheme inflated gasoline prices by up to 22 cents per gallon and diesel by up to 33 cents per gallon, per Insurance Journal. The plaintiffs — California drivers — are seeking damages for the antitrust violations, though the total damages figure has not been publicly specified in available reporting.

The core legal theory tracks what antitrust practitioners call algorithmic collusion: the idea that competing firms, without explicit human-to-human coordination, can use a shared or functionally equivalent pricing algorithm to achieve supracompetitive prices. Courts and regulators have wrestled with this theory for years, and this filing is among the more prominent retail-fuel applications of it in the United States.

The defendant list is notable for its breadth. It spans vertically integrated oil majors (BP, Marathon Petroleum), convenience-store and fuel-retail chains (7-Eleven, Speedway), and general merchandise retailers with fuel operations (Walmart, Albertsons). That mix suggests the plaintiffs' theory does not depend on a single shared software vendor but rather on parallel adoption of AI-driven dynamic pricing tools that, in aggregate, produced price alignment across nominally competing outlets.

Algorithmic pricing in fuel retail is not new. Stations have used automated systems to track competitor prices and adjust their own in near-real time for well over a decade. What has shifted is the sophistication of those tools — machine-learning models can now optimize margins across thousands of variables simultaneously and respond to competitor moves in seconds rather than minutes. The legal question is whether that speed and coordination, even absent explicit agreement, constitutes a per se or rule-of-reason violation under the Sherman Act or California's Cartwright Act.

That question has no settled answer. The Department of Justice and FTC have flagged algorithmic collusion as a priority enforcement area, and academic literature — most influentially work by Ariel Ezrachi and Maurice Stucke — has long argued that existing antitrust frameworks struggle to catch tacit machine coordination. But no U.S. court has yet issued a definitive ruling on whether competing firms independently deploying functionally similar AI pricing tools constitutes an unlawful agreement. This case, if it survives early motions to dismiss, could become a significant test of that boundary.

The named defendants are also well-resourced litigants. Expect vigorous challenges on the "agreement" element — antitrust plaintiffs must prove concerted action, not merely parallel conduct, and defendants will argue that price similarity in a transparent commodity market is consistent with independent, lawful competition. Discovery into the specific AI systems used by each company will be the pivotal battleground: if plaintiffs can show defendants shared a common platform or fed data into an interoperable network, the agreement element becomes considerably easier to plead.

California is a particularly consequential venue. The state's gasoline prices are among the highest in the country, amplifying per-gallon damages claims, and its consumer protection statutes offer plaintiffs tools that federal law alone does not. The Cartwright Act, California's analog to the Sherman Act, has been read broadly by state courts and carries per se treatment for horizontal price-fixing — a framing the plaintiffs will likely push hard.

The filing lands as AI pricing tools are proliferating across retail sectors well beyond fuel. Grocery, airline, and hospitality industries all deploy dynamic pricing at scale, and the antitrust exposure question is live in each of them. A ruling — or even a significant settlement — in this case would send a pricing signal of its own to every legal and compliance team managing algorithmic revenue optimization.

None of the named defendants had publicly responded to the complaint as of the reporting reviewed for this piece.