A class-action filed in Sacramento federal court names BP, Marathon, 7-Eleven, Walmart and Albertsons, alleging their shared use of Kalibrate Fuel Systems' pricing algorithm inflated California gas prices by as much as 22 cents a gallon.
California already has the most expensive gasoline in the country. According to the complaint filed June 22 in Sacramento federal court, plaintiffs say part of the reason sits inside a software platform. The lawsuit alleges that BP, Marathon Petroleum, 7-Eleven, Walmart, Albertsons and Circle K, companies that collectively operate more than 1,700 filling stations across the state, fed confidential pricing data into an algorithm built by Kalibrate Fuel Systems. That algorithm, the plaintiffs say, then recommended near-identical prices across competing outlets, pushing gasoline up by as much as 22 cents a gallon and diesel by 33 cents above where a competitive market would have landed them. In some parts of the state, Bloomberg reported, prices had already crested $7 a gallon, partly because of fuel supply disruptions tied to the U.S. conflict with Iran.
The numbers are not abstract. Every additional penny per gallon costs California drivers roughly $134 million a year, according to the complaint. At 22 cents, that's close to $3 billion annually, extracted from motorists by what plaintiffs describe as a hub-and-spoke conspiracy: Kalibrate as the hub, the competing retailers as spokes, each one separately handing pricing decisions to the same system without needing to call a rival.
That legal theory is the thing to watch. The DOJ's antitrust division has already argued, most prominently in its ongoing litigation against RealPage, that competitors routing pricing decisions through a shared algorithm can amount to collusion even when they never speak directly. RealPage's software collected rent data from competing landlords and used it to recommend prices that prosecutors alleged kept rents artificially high across whole housing markets. The California gas case follows a similar path, only this time you're not looking at rent checks. You're looking at the price posted on a sign drivers pass every week.
The lawsuit is also among the first actions filed under California's Assembly Bill 325, which took effect January 1, 2026. AB 325 is the strongest state-level law in the United States aimed directly at algorithmic price-fixing. It prohibits shared pricing algorithms when competitors feed data into the same system and receive coordinated price recommendations in return. The Cartwright Act, California's main antitrust statute, gives the complaint its other legal backbone.
Don't underestimate the software question here. The unresolved issue is whether Kalibrate shares exposure alongside the fuel retailers, or whether legal risk sits only with the companies that chose to use the platform. A ruling that reaches the vendor changes the calculation for enterprise AI procurement. Any founder selling dynamic pricing software into retail, logistics or SaaS needs to read this complaint carefully, because the operators here presumably believed they were buying a lawful optimization tool. The court now has to decide when efficient pricing technology becomes facilitated collusion.
Kalibrate is not a marginal vendor. The company markets itself as a global provider of fuel pricing intelligence, and the Canadian Competition Bureau has already opened an investigation into its pricing services in Canada to determine whether its analytics were helping fuel retailers coordinate prices there. That investigation predates the California filing. If the Sacramento case produces a finding against the platform itself, the consequences would run well beyond the 1,700 stations named in this complaint.
For now, the defendants haven't publicly responded to the suit. BP, Marathon and Walmart did not immediately comment, according to Bloomberg's reporting. Kalibrate also declined to comment. That silence is normal, but the filing is detailed enough that the central fight is already visible: plaintiffs will argue that sharing confidential pricing data with a common algorithm is coordination; defendants will argue they used a commercially available tool while making independent business decisions. Courts haven't settled that question. This case gives them a clean chance to do it.
Here's the thing for anyone building or buying AI pricing tools: the era of the algorithm as a liability shield is ending. Regulators and plaintiffs are getting more comfortable with the hub-and-spoke theory, and California just gave them a statute written for this exact problem. If your pricing model depends on competitors sending confidential data into the same platform and getting price recommendations back, you don't get to wave away the antitrust risk by calling it software.
Also read: OpenAI's Patch the Planet program takes its security ambitions well beyond the chatbot business • AI Financial Modeling for Startups Is the CFO Alternative That Actually Works • Google lost two of its most important AI researchers within days and the market wiped out $250 billion to make the point
