AI pricing tools have moved from hotel rooms and apartments to the drive-thru. On October 2, 2026, a customer filed a proposed nationwide class action in Chicago. It claims McDonald's used a machine-learning pricing engine to coordinate menu prices across its own restaurants and thousands of independent franchisees.

McDonald's denies it and says AI does not set the price of any menu item. Nothing has been proven yet. But the case is a clear warning for anyone who builds pricing software. I read the complaint, the Reuters coverage and McDonald's response. Here is what agencies in India and the USA should take from it.

McDonald's digital drive-thru menu board showing burger meal prices outside a restaurant in Virginia — A digital menu board at a McDonald's in Ashburn, Virginia. Photo: Julie Littman/Restaurant Dive

What the AI Pricing Lawsuit Against McDonald's Claims

The case is Thomas v. McDonald's USA, LLC and McDonald's Corporation. It was filed in the US District Court for the Northern District of Illinois. The plaintiff, Michael Thomas, is an Illinois resident. He wants to represent every person in the US who bought food at a McDonald's, plus an Illinois subclass.

As Reuters reported on the McDonald's class action, the suit says the company broke US antitrust law by working with franchisees to fix prices using algorithms trained on nonpublic data. The complaint lists four counts:

Two counts under Section 1 of the Sherman Act, the main US law against price fixing.
One count under the Illinois Antitrust Act.
One count under the Illinois Consumer Fraud and Deceptive Business Practices Act.

The plaintiff is asking for treble damages, which means three times the actual losses if he wins.

The core allegations

The complaint makes several specific claims. Keep in mind these are allegations, not findings:

A central pricing engine since at least 2019. It allegedly takes in data from millions of daily transactions across corporate and franchised stores.
Franchisee data is pooled. About 95 percent of the roughly 14,000 US restaurants are independently owned. The suit says their confidential sales data trains the same model.
Willingness to pay. The engine allegedly estimates how much each store's customers will accept, not just what covers costs.
Linked price rises. The suit says the algorithm has rules that tie one store's price increase to whether other McDonald's stores already raised prices on the same item.
Pressure to comply. It claims that since January 2026, franchisees must use the company's approved pricing consultant and tools, and that deviations are tracked.

Where the 40 percent figure comes from

You will see a 40 percent number in the coverage. The complaint takes it from a McDonald's fact sheet dated May 29, 2024. That fact sheet said average menu prices rose about 40 percent between 2019 and 2024. The lawsuit argues the pricing tool helped drive that rise. McDonald's has blamed inflation and rising costs in the past.

The complaint also points to a viral $18 Big Mac meal at a franchise in Darien, Connecticut. That owner, George Michell, sued McDonald's separately in 2024 in a dispute over his franchise.

How the Pricing Tool Reportedly Works

The lawsuit leans heavily on a Reuters investigation published on September 29, 2026. That report said McDonald's pricing engine uses machine-learning algorithms to keep analysing data from its nearly 14,000 restaurants.

According to the complaint, which quotes Reuters, the tool measures "customer willingness to pay in your area." It also warns owners when their prices look too low, with messages like "Your restaurant is showing MEDIUM SENSITIVITY to Price."

One detail stands out for software teams. The portal's terms reportedly warn that users "may be competitors of each other" and must follow antitrust law. Former FTC Commissioner William Kovacic told Reuters that this disclaimer was "an acknowledgment there's a potential problem."

Reuters also reported that five store owners said the company pressured them to use the tools. A June franchisee document reviewed by Reuters showed that deviations from the recommendations were recorded in detail.

First page of the class action complaint Thomas v. McDonald's USA filed in the Northern District of Illinois — First page of the class action complaint filed on October 2, 2026. Source: US District Court, N.D. Illinois, via Reuters

McDonald's Response to the AI Pricing Claims

McDonald's called the allegations "speculative and uninformed" in a statement on October 5. "AI does not set the price of a Big Mac or any other menu item," the company said. It added that franchisees make their own pricing decisions and that pricing recommendation tools are common across industries.

The company made similar points after the Reuters report. Restaurant Dive noted that McDonald's said it does not use dynamic pricing. It said its recommendations are restaurant-specific and do not "change prices in real-time or at different hours during the day."

That distinction matters. Dynamic pricing changes prices by time or demand, like surge fares. The lawsuit is about something else: a shared model that may line up prices between businesses that are supposed to compete.

There is also wider context. In September, McDonald's announced an $8.5 billion partner investment through 2036 and an AI restaurant system called ArchIQ. In 2024, Wendy's dropped plans to test dynamic pricing after public backlash. Customers clearly notice when software touches what they pay.

Why AI Pricing Is Now an Antitrust Issue

This is not a one-off. Reuters noted that US plaintiffs have filed a wave of class actions claiming companies used algorithms to coordinate prices for hotel rooms, apartment rentals and other purchases.

The legal idea behind many of these cases is often called hub-and-spoke collusion. Competitors never talk to each other directly. Instead, they all feed private data into one shared tool, and the tool becomes the hub.

The complaint quotes a March 2025 statement of interest from the US Department of Justice. In it, the DOJ argued that a formula used to fix recommended or starting point prices can break the Sherman Act even when final prices differ. It also said that sharing data through an algorithm provider can have the same effect as competitors swapping it directly.

California now names the problem in law

California went further. California's AB 325 on common pricing algorithms was approved on October 6, 2025, and took effect on January 1, 2026. It defines a common pricing algorithm as any software "used by two or more persons, that uses competitor data to recommend, align, stabilize, set, or otherwise influence a price or commercial term."

The law makes two things unlawful:

Using or distributing such an algorithm as part of a conspiracy to restrain trade.
Using or distributing one while coercing someone to adopt its recommended price.

The McDonald's suit was filed under federal and Illinois law, not California law. Still, its claims about pressure on franchisees map closely to the coercion clause. If you sell pricing software to clients with California customers, read that section carefully.

India's competition regulator is watching too

Indian teams should not assume this is only an American story. The Competition Commission of India's market study on AI and competition, carried out through MDI Gurgaon, flags algorithmic collusion as an emerging risk.

The study describes monitoring algorithms, hub-and-spoke setups, signalling algorithms and self-learning systems that can reach collusive outcomes on their own. Its advice to businesses is practical. It urges self-audits of AI systems for competition compliance and plain-language transparency about how algorithms affect prices.

What AI Pricing Risk Means for Agencies in India and the USA

Most agencies will never build a pricing engine for 14,000 restaurants. But many of us build the pieces. Think retail dashboards, ecommerce price rules, franchise portals, SaaS products for landlords or clinics, and recommendation models that suggest prices.

The risk is rarely in one feature. It sits in how data and control flow between clients who compete. Our guide to AI liability risks for agencies covers the wider question of who gets blamed when software causes harm. Pricing is one of the sharpest versions of that problem.

Red flags in a pricing project brief

If a client brief includes any of these, slow down and involve their legal team:

One model trained on nonpublic sales data from competing businesses.
Rules that raise a price because a competitor already raised theirs.
Dashboards that score or report on users who ignore the recommended price.
Any link between following recommendations and contract renewal, ranking or access.
"Willingness to pay" targeting at the level of an individual shopper with no clear disclosure.

Safer design choices compared

Feature	Higher-risk version	Safer version
Training data	Pooled private data from competing clients	Each client's own data plus public market data
Recommendations	Tied to rivals' recent price moves	Based on the client's own costs and demand
Overrides	Tracked and reported to a central body	Logged privately for the business that made them
Final price	Applied automatically across locations	Approved by a person with a recorded reason
Customer view	No explanation of how prices are set	Clear notice when prices are personalised

A Practical AI Pricing Checklist for Client Builds

Here is the checklist we now use before a pricing feature goes into a sprint:

Map the data. List every data source behind each price suggestion and who owns it.
Isolate tenants. In multi-client SaaS, keep one client's sales data out of another client's model.
Keep a human in charge. Let people set final prices, and log who approved what and why.
Do not punish overrides. Avoid compliance scores or alerts that push users toward the suggested price.
Write it down. Document model inputs, rules and tests, as the CCI study recommends for self-audits.
Test for alignment. Check whether prices across separate clients move together after launch.
Get legal sign-off. Have the client's counsel review anything that touches competitor data.

None of this replaces legal advice. It does make the right questions show up early, when changes are cheap. If you want a second pair of eyes on an existing system, an AI readiness assessment is a good place to start.

What This Looks Like on a Real Project

Picture a franchise food brand with 200 outlets across Delhi, Mumbai and Bengaluru. It asks for a central dashboard that suggests menu prices for every outlet. Each outlet is owned by a different franchisee.

The easy build pools every outlet's sales into one model and pushes the same suggestions to all of them. After this lawsuit, we would push back. We would train on each outlet's own data, show public market signals separately and never report who ignored a suggestion to head office.

The same logic applies to a US client running a marketplace for independent sellers. Shared tools are fine. Shared private data that steers competitors toward the same price is where the trouble starts. Our software development consulting team now asks these questions during discovery, not after launch.

What Comes Next

The McDonald's case is at the very first stage. A judge has not ruled on anything, and McDonald's will likely move to dismiss. The outcome could take years.

For builders, the direction is already clear. Regulators in the US and India are looking at shared pricing algorithms, and plaintiffs are filing more of these suits. Building with clean data boundaries now costs far less than rebuilding later.

Key Takeaways

A proposed class action claims McDonald's used an AI pricing engine to coordinate menu prices with franchisees. McDonald's denies it.
The core concern is pooled nonpublic data and rules that link one seller's price to another's.
California's AB 325 already bans coercive use of common pricing algorithms.
India's CCI recommends self-audits and transparency for AI that influences prices.
Agencies should isolate client data, keep humans in charge of final prices and get legal review early.

About the Author

Nitin Kumar is the Lead Project Coordinator at Debugged Pro. He coordinates development and QA teams delivering web platforms, ecommerce systems and AI integrations for clients across India and the USA, and he tracks AI regulation so client builds stay on the right side of the law.