McDonald's AI Pricing Lawsuit Tests the Line Between Advice and Price Fixing

McDonald’s faces a proposed nationwide class action over an AI-assisted pricing system that allegedly coordinates menu prices across thousands of independently operated restaurants.

The McDonald’s AI pricing lawsuit was filed on October 2, 2026, in federal court in Chicago. Plaintiff Michael Thomas claims the company pooled confidential restaurant data and converted it into pricing recommendations that suppressed competition among franchisees.

McDonald’s disputes that account. It says artificial intelligence does not set menu prices and that restaurant operators remain free to accept or reject its recommendations. That disagreement creates the central question: when does centralized pricing advice become coordination among businesses that are supposed to compete?

The complaint has not established wrongdoing, and no court has certified the proposed class. However, the dispute arrives after federal antitrust agencies pursued algorithmic pricing cases in housing and examined data-driven pricing across consumer markets.

The case is therefore larger than the price of a Big Mac. It tests whether a franchisor can combine data from nominally independent restaurants without turning a management tool into a coordination mechanism.

What the McDonald's AI Pricing Lawsuit Alleges

The complaint attacks the data-sharing arrangement behind the recommendations, not merely the presence of artificial intelligence.

Thomas filed the case against McDonald’s USA and McDonald’s Corporation in the U.S. District Court for the Northern District of Illinois. The public federal docket identifies it as Thomas v. McDonalds USA, LLC, case number 1:26-cv-12149.

Thomas lives in DeKalb, Illinois, and says he regularly buys meals at McDonald’s restaurants near his home. According to the complaint, he observed different prices for identical items at nearby locations.

Price differences alone do not prove an antitrust violation. Independent businesses ordinarily can charge different prices based on rent, wages, local demand, and operating expenses.

The complaint makes a different argument. It alleges that McDonald’s created a common pricing engine that collects current, nonpublic information from restaurants before recommending menu prices to those same restaurants.

That distinction matters because about 95 percent of McDonald’s roughly 14,000 U.S. restaurants are independently owned and operated. Those franchisees are separate legal businesses, even though they trade under one brand.

The filing contends that participating restaurants supply granular sales and pricing information to a shared system. The system allegedly analyzes millions of daily transactions, store characteristics, local demand, customer price sensitivity, and competitor prices.

It then generates restaurant-specific recommendations for individual menu items. The complaint describes those outputs as part of a coordinated system rather than neutral market research.

Thomas also alleges that the engine applies a “30% rule.” Under the complaint’s account, the tool can recommend an increase after a specified share of nearby restaurants has raised the same item’s price.

That allegation has not been tested through discovery. McDonald’s has not accepted the plaintiff’s description of the tool or its operating rules.

The complaint further claims that McDonald’s monitors whether franchisees follow its recommendations. It alleges that persistent deviations can affect how the company evaluates an operator’s eligibility for expansion or franchise renewal.

This pressure allegation is essential to the plaintiff’s theory. A recommendation that operators routinely ignore looks different from a recommendation backed by monitoring and commercial consequences.

The proposed class includes U.S. consumers who purchased food from McDonald’s during the alleged period. Thomas also seeks to represent a subclass of Illinois customers.

The complaint asserts claims under Section 1 of the Sherman Act, the Illinois Antitrust Act, and the Illinois Consumer Fraud and Deceptive Business Practices Act. It seeks class certification, damages, legal fees, and an injunction against the alleged arrangement.

Under federal antitrust law, successful private plaintiffs can seek treble damages, meaning three times their proven losses. That remedy does not mean damages have already been established.

The complaint remains an initiating allegation. No judge has decided whether the restaurants agreed to coordinate, whether the system restrained competition, or whether customers suffered measurable overcharges.

Those unresolved questions separate a serious legal claim from a completed finding of price fixing.

McDonald's Says AI Does Not Set Menu Prices

McDonald’s defense rests on operator independence: recommendations provide context, while each franchisee makes the final pricing decision.

Before the lawsuit was filed, McDonald’s published a detailed response to reporting about its pricing technology. Its October 1 pricing statement says AI does not determine the price of any menu item.

The company describes its system as a restaurant-specific recommendation tool. It says operators use those recommendations alongside their own knowledge of local costs, customers, and competition.

McDonald’s also rejects the idea that it uses real-time dynamic pricing. Dynamic pricing generally changes prices quickly in response to immediate conditions, such as demand, inventory, or time of day.

According to the company, its tool does not automatically change digital menu boards during busy periods. It also does not assign different prices to individual customers based on personal information.

That distinction separates three practices that public discussions often combine.

Algorithmic pricing uses software to analyze information and recommend or select prices. Dynamic pricing changes a price as market conditions change. Surveillance pricing uses personal or behavioral data to tailor prices or promotions for particular consumers or groups.

The complaint primarily concerns algorithmic coordination among restaurants. It does not establish that McDonald’s charges two people at one location different prices for the same order.

McDonald’s told the Associated Press that the complaint is filled with inaccuracies and that it will defend itself. The company said its optional tools neither automate nor coordinate restaurant prices.

A company spokesperson also emphasized that McDonald’s was collecting market information and recommending prices before machine learning entered the process. From that perspective, AI improves an established advisory function.

Franchisee behavior provides some support for the claim that corporate recommendations are not automatic commands. During an August investor call, CEO Chris Kempczinski said only 60 percent of U.S. restaurants offered a proposed value menu.

He said McDonald’s could not simply activate the offer across the system. Corporate management had to reach an agreement with franchisees.

That example does not resolve the lawsuit. A franchisee might reject a national promotion while still relying heavily on restaurant-specific pricing recommendations.

It does show why the dispute cannot be decided by treating every McDonald’s location as a corporate branch. The franchise system divides authority between the brand owner and thousands of local operators.

McDonald’s also has legitimate reasons to provide pricing analysis. Franchisees need to account for food costs, labor expenses, local competitors, customer traffic, and the effect of price changes on sales.

A recommendation system can process more transactions than an individual operator could evaluate manually. It might also identify when a higher price would reduce customer traffic enough to lower total revenue.

The legal issue is not whether software can produce useful analysis. It is whether McDonald’s built that analysis from competitively sensitive inputs and encouraged nominal rivals to follow a common output.

Final pricing discretion helps McDonald’s defense, but it is not necessarily decisive. Courts can examine how a system operates in practice, including participation, monitoring, pressure, and adoption rates.

The company’s strongest evidence would show meaningful independent decision-making. That could include frequent departures from recommendations, varied pricing methods, and an absence of penalties tied to compliance.

The plaintiff, by contrast, needs evidence that participating restaurants understood their competitors were contributing data and expected others to follow the same system.

Those facts are unlikely to emerge from public statements alone. They would require internal documents, operator communications, technical records, and testimony about how recommendations were implemented.

Why Shared Franchise Data Creates the Antitrust Risk

The core conflict is independent pricing versus a common engine trained on information that competitors would not ordinarily exchange.

Section 1 of the Sherman Act prohibits agreements that unreasonably restrain trade. Traditional price fixing involves competitors agreeing directly on the prices they will charge.

Software does not create a legal exemption. If competitors cannot lawfully coordinate through a meeting, they generally cannot make the same arrangement lawful by routing decisions through an algorithm.

The Justice Department has taken that position in several recent disputes. In a hotel pricing case, the department and Federal Trade Commission said companies cannot use algorithms to perform conduct that would be unlawful if performed by people.

Their hotel pricing filing also addressed the role of recommendations. The agencies argued that a common starting price can create antitrust concerns even when businesses retain some discretion.

That view supports the legal theory behind the McDonald’s complaint. It does not prove that McDonald’s operated such an agreement.

The franchise structure adds a difficult boundary question. McDonald’s owns the brand and establishes standards for products, restaurant operations, marketing, and customer experience.

Uniformity is central to a franchise. Customers expect similar food, service, and branding across locations, even when different companies own the restaurants.

Prices occupy a more sensitive category. McDonald’s has repeatedly stated that independent operators set their own menu prices and compete for customers in local markets.

If those operators truly compete, their private sales data can reveal how customers responded to particular price changes. Pooling that information can reduce uncertainty about what rivals plan to charge.

Competitive uncertainty is not a defect that every optimization tool should remove. It is one of the forces that encourages businesses to lower prices, improve service, or accept smaller margins.

The complaint alleges that McDonald’s reduced that uncertainty by collecting store-level information and returning coordinated recommendations. It portrays the company as the central hub connecting otherwise independent operators.

Antitrust lawyers often call this a hub-and-spoke arrangement. A central actor serves as the hub, while participating competitors form the spokes.

A hub-and-spoke theory still requires evidence of an agreement. It is not enough to show that multiple businesses purchased the same software or received similar advice.

The plaintiff must connect the vertical relationships between McDonald’s and each franchisee to a horizontal understanding among competing restaurants. Courts examine whether participants knew others were joining and expected coordinated conduct.

The role of company-operated restaurants adds another layer. McDonald’s subsidiaries operate a smaller share of U.S. locations, potentially placing corporate restaurants in direct competition with franchisees.

If corporate stores contributed data and received recommendations from the same engine, the plaintiff will argue that McDonald’s participated as both platform operator and restaurant competitor.

McDonald’s can respond that shared branding and franchise support naturally require centralized analysis. It can also argue that restaurant-level recommendations reflect local differences rather than suppressing them.

Visible price variation supports that point. Thomas himself noticed different prices at nearby stores, which suggests that the system did not impose a single national menu price.

However, antitrust coordination does not always produce identical prices. A common system can recommend different prices while still reducing independent decision-making.

The relevant question is therefore counterfactual: what prices would restaurants have selected without the pooled data, common rules, and alleged pressure?

That question makes the McDonald’s AI pricing lawsuit an economic case as much as a technology case. The parties will need to separate coordinated effects from inflation, wages, ingredient costs, rent, and shifting consumer demand.

Higher Menu Prices Do Not Prove Algorithmic Price Fixing

McDonald’s prices rose sharply, but a correlation between the pricing tool and those increases cannot establish an unlawful agreement by itself.

The complaint points to McDonald’s own statement that average U.S. menu prices rose about 40 percent between 2019 and 2024. It links that period to the development and expansion of the company’s pricing technology.

The timing creates a plausible research question. It does not isolate the algorithm’s effect.

Restaurants faced major increases in labor, food, packaging, transportation, and occupancy costs during that period. Consumer behavior also changed during the pandemic and the inflation that followed.

McDonald’s previously said menu prices increased partly to offset higher business expenses. Franchisees in different regions encountered different wage laws, rents, taxes, and supply conditions.

A reliable damages model would need to control for those factors. It would also need a benchmark showing how comparable prices moved without the challenged system.

Nearby McDonald’s locations offer one possible comparison, but local ownership can complicate it. The same franchisee may own several restaurants within a region.

The proposed class also covers a large and varied market. A tool could influence some restaurants strongly while having little effect on others.

The plaintiff will need evidence that the alleged coordination produced prices above competitive levels across a legally coherent class. That can become difficult when recommendations, adoption, costs, and ownership differ by location.

The underlying reporting identified striking local price gaps. One example involved two Fresno restaurants located about two miles apart, where listed Big Mac prices reportedly differed by 21 percent.

Such variation can support competing interpretations. It might show independent pricing, or it might show that a centralized engine optimizes the amount each local market can bear.

Machine learning itself does not answer that question. It is a method for finding patterns in data and producing predictions or recommendations based on those patterns.

A model that estimates willingness to pay can recommend different prices for different locations without identifying individual customers. That practice resembles localized price optimization rather than personalized surveillance pricing.

Localized pricing is not automatically unlawful. Businesses routinely adjust prices for regional costs, demand, and competitive conditions.

The risk changes when the model incorporates current, confidential data from businesses that should be making independent decisions. The source and timing of the inputs may matter more than the sophistication of the model.

The FTC pricing study shows why regulators increasingly examine data pipelines. Its initial findings described intermediaries capable of using location, browsing activity, demographics, and shopping behavior to influence prices or promotions.

The FTC’s study concerns a broader category than the McDonald’s allegations. It does not identify McDonald’s as having charged individualized prices.

However, it reflects a shift in regulatory attention. Agencies increasingly ask what data enters pricing systems, who shares it, and whether customers can understand the resulting offers.

The complaint will face another challenge because much of its public narrative relies on reporting and company materials. Litigation requires admissible evidence that supports each legal element.

Discovery could reveal a coercive system built to align operators. It could instead show optional recommendations, weak adoption, and substantial price variation driven by local conditions.

Until those records emerge, claims that McDonald’s “used AI to raise everyone’s price” go beyond the verified facts. The accurate statement is narrower: a consumer alleges that the pricing system enabled unlawful coordination.

The Real Precedent Comes From Housing and Hotels

This lawsuit brings an established algorithmic-collusion theory into a franchise network, where the relationship among participants is less straightforward.

The most prominent recent algorithmic pricing cases have involved landlords, hotels, and software intermediaries. Those cases generally allege that competing firms supplied confidential data to a common vendor and received pricing recommendations.

In 2024, the Justice Department sued RealPage, alleging that its revenue management system helped competing landlords coordinate apartment rents. The government later added several major landlords as defendants.

The department’s RealPage case described the software as a mechanism for sharing sensitive information and aligning rental decisions. RealPage disputed the allegations.

That litigation gives the McDonald’s plaintiff a vocabulary and legal roadmap. Both theories focus on pooled nonpublic data, common recommendations, and participants who allegedly surrendered independent judgment.

The differences are substantial. RealPage served separate property companies that purchased software from an outside vendor.

McDonald’s operates a franchise network with shared trademarks, supply standards, advertising, products, and contractual relationships. Corporate coordination is built into many lawful parts of that system.

A court must distinguish brand management from horizontal price coordination. That makes operator independence the decisive issue.

Hotel pricing litigation presents a related precedent. Hotel owners often participate in the same branded network while remaining independent businesses.

Federal agencies have argued that voluntary recommendations can still support a price-fixing claim when participants knowingly use a common pricing system. Courts, however, continue to evaluate these cases under their specific facts.

The McDonald’s dispute might also influence other franchise industries. Restaurant, hotel, fitness, automotive, and service brands increasingly supply operators with centralized analytics.

Those tools can improve forecasting, reduce waste, and help small operators interpret local demand. They can also create a detailed view of competitor behavior that no participant could obtain independently.

The safest distinction is not “AI versus no AI.” Traditional spreadsheets, consultants, or phone calls can facilitate the same unlawful conduct.

A better compliance question asks whether the system preserves genuine independence. Relevant safeguards include aggregated historical data, limits on competitor information, voluntary participation, and restrictions on monitoring adoption.

Real-time, store-level information presents greater risk because it can reveal current competitive strategy. Recommendations linked to rivals’ recent decisions also invite closer scrutiny.

Businesses using algorithmic pricing should therefore document more than a final-choice disclaimer. They need operational evidence showing that customers retain control and that the provider does not punish deviations.

Franchise systems face an added governance problem. A recommendation may be technically optional while carrying practical weight because the franchisor controls renewals, expansion opportunities, technology access, and brand standards.

That does not make every recommendation coercive. It means the surrounding business relationship can matter as much as the software interface.

The eventual importance of Thomas v. McDonald’s will depend on how the court treats that relationship. A broad ruling could expose many centralized franchise tools to new claims.

A narrow ruling could confirm that franchisors may provide sophisticated pricing guidance when operators remain independent and confidential competitor data is properly protected.

For now, the complaint is a pressure test, not a precedent.

What to Watch as the McDonald's AI Pricing Lawsuit Develops

Three signals will show whether this case becomes a major algorithmic antitrust precedent or ends as an unsuccessful challenge to franchise support tools.

The first signal is McDonald’s initial motion practice. The company will likely challenge whether the complaint plausibly alleges a horizontal agreement among franchisees.

A dismissal ruling would reveal how much factual detail a consumer must provide before gaining access to internal records. Survival would allow discovery into the pricing engine and franchise relationships.

The key issue will not be whether the complaint repeatedly uses the term AI. It will be whether the alleged system connects competitors through a shared commitment to follow common recommendations.

The second signal is evidence about adoption and enforcement. Internal documents may show how often operators accepted recommendations, how McDonald’s tracked deviations, and what happened when an operator declined.

High adoption would not independently prove an agreement. Yet high adoption combined with pressure, competitor awareness, and nonpublic inputs would strengthen the plaintiff’s theory.

Frequent deviation without consequences would strengthen McDonald’s position. It would support the company’s claim that the tool supplies context rather than commands.

The third signal is the court’s treatment of data. The complaint emphasizes current, granular information from individual restaurants because that information can expose competitive strategy.

A system trained primarily on aggregated or historical information presents a different risk from one using live store-level sales and prices. Technical discovery should clarify which description fits the McDonald’s engine.

Readers should also watch whether regulators participate. The Justice Department and FTC have filed statements in private algorithmic pricing cases when broader interpretations of antitrust law were at stake.

An agency filing would not determine the result. It would signal that federal enforcers view the franchise context as important to algorithmic competition policy.

The case also arrives as McDonald’s tries to restore its value reputation among budget-conscious customers. That business pressure gives the dispute consequences beyond legal damages.

If customers believe software is designed to identify the highest tolerable price, corporate messaging about affordability becomes harder to sustain. McDonald’s must explain both how the tool works and what limits govern its use.

Developers and enterprise buyers should pay attention for a different reason. The case shows that a technically advisory product can create legal risk through its inputs, incentives, and deployment model.

Product teams should ask whether their systems combine confidential data from competitors. They should also examine whether recommendations become effectively mandatory through monitoring, contractual leverage, or performance reviews.

Consumers should avoid assuming that every location-based price difference reflects misconduct. They should also expect clearer explanations when companies use shared data to guide prices across independently operated businesses.

The McDonald’s AI pricing lawsuit will turn on evidence that is not yet public. The complaint describes coordinated price fixing, while McDonald’s describes optional business advice.

Watch the dismissal briefing, recommendation adoption records, and technical data controls. Together, those signals will show whether the engine preserved competition or quietly replaced it.