If you have ever paid for a Big Mac and thought "this was cheaper last time", you may not have been imagining it. You may simply have been at a different McDonald's.
In September, Reuters checked two company-run McDonald's restaurants in Fresno, California, about two miles apart. One charged $5.69 for a Big Mac. The other charged $6.89. Same burger, same company, same city: a $1.20 gap, or 21%.
Behind that gap, according to Reuters' investigation, sits a machine-learning pricing engine that McDonald's has used across its US restaurants since at least 2019. It looks at millions of daily transactions, at competitors' menus, and at one input it reportedly weighs heavily: an estimate of how much the customers of each individual restaurant are willing to pay.
Reuters could not prove the engine alone caused the Fresno gap. Local costs, ownership and promotions also matter. But the system exists, it covers nearly 14,000 restaurants, and it changes how you should think about the price on the menu board. This is what it means for your wallet.
How the pricing engine works
The description below comes from Reuters' reporting, McDonald's own statements and the documents franchisees showed reporters.
| What the system does | What it means for you |
|---|---|
| Analyzes millions of daily transactions | it knows what sells at your store, at what price, and at what time |
| Pulls menus from nearby Wendy's and Burger King | your price partly depends on what the competitor down the road charges |
| Estimates each store's "price sensitivity" | if your store's customers keep buying when prices rise, it suggests charging more |
| Recommends an "optimal price" for every item at every store | two restaurants in one town can get different recommendations |
| Franchisees can accept, change or reject it | the final price is set by the owner, but under some pressure, as below |
One detail from Reuters stands out. Operators received ratings such as: "your restaurant is showing MEDIUM SENSITIVITY to price." Put simply, the system grades each restaurant by how much its customers push back when prices go up. A store whose customers rarely push back gets nudged towards higher prices.
That is the key difference from what most people imagine when they hear "AI pricing". This system is not reading your phone or your profile. It is pricing the neighbourhood: the commuters, the tourists, the people with no other option nearby. If you are one of them, you pay the neighbourhood's price.
"A tool, not a mandate"? What franchisees say
McDonald's says the engine is "a tool, not a mandate, designed to provide restaurant-specific recommendations", and that franchisees, who run the large majority of US restaurants, keep the final say on prices.
The franchisees Reuters spoke to described something less relaxed:
- 5 owners said they felt pressured to use the tools.
- 3 said the engine had widened price gaps between neighbouring restaurants.
- Internal standards reportedly require franchisees to be "constructively engaging with McDonald's approved Pricing Consultant and Tools", and corporate tracks how often they follow the advice.
- On an earnings call, CEO Chris Kempczinski said that "pricing non-compliance in certain cases is part of those conversations" when franchise renewals come up. McDonald's controls renewals and approvals for new locations, so that sentence carries weight.
The best-known extreme case predates this investigation: a Big Mac meal priced around $18 at a Connecticut turnpike location in 2024. McDonald's said at the time it was a single restaurant out of more than 13,700. Coverage of a franchisee's lawsuit has since claimed the engine itself suggested that price. McDonald's disputes the lawsuit, and Reuters could not verify the claim.
How much have prices actually risen?
This is the backdrop that makes the pricing engine sting. By McDonald's own figures, published in 2024 by its US president:
| Item | 2019 average | 2024 average | Increase |
| Big Mac | $4.39 | $5.29 | +21% |
| Big Mac meal | $7.29 | $9.29 | +27% |
| 10-piece McNuggets meal | +28% | ||
| Medium fries | +44% | ||
| Average menu item | about +40% |
Independent surveys claim far bigger rises, and McDonald's disputes them. Either way, customers noticed. Kempczinski told investors that visits from lower-income customers had fallen by nearly double digits over two years.
Now compare the two Fresno prices with the company's own average. $5.69 is a little above the 2024 national average. $6.89 is 30% above it. That second number is the one the price-sensitive grading produces when it meets a captive audience.
The money math for an ordinary family
The gap per item looks small. It adds up because fast food is a habit.
Take a family of four who eat at McDonald's once a week and order four main items. Assume they usually use a store that charges $1 more per item than a comparable restaurant a few minutes away. That is roughly the Fresno gap.
| Per visit | Per year (52 visits) | |
| Extra on 4 items | $4 | $208 |
| If drinks and fries carry the same premium (8 items) | $8 | $416 |
None of that is a fortune. But it is money paid for nothing at all: the same burger, from the same kitchen procedures, two miles apart. And the same logic is spreading to the other places you spend every week.
It is not only McDonald's
Groceries. A study by Consumer Reports, Groundwork Collaborative and More Perfect Union had 437 shoppers in four cities add identical items to Instacart at the same moment. About three in four items showed more than one price. The average gap between the lowest and highest price was 13%, up to 23%. The authors estimated that a typical family of four could pay around $1,200 a year more. Instacart disputed the figure, said these were randomized tests run with retailers, and has since ended item price tests.
Other chains. Wendy's faced a backlash in 2024 after its CEO talked about dynamic pricing, and said the comments were misunderstood. Walmart's CEO has said the company will not price by shopper identity or time of day. Wendy's and Burger King told Reuters they do not use AI in pricing decisions.
The wider picture, including prices tailored to you personally from your data, is in how surveillance pricing is quietly costing you. The McDonald's case is the milder cousin: it prices places, not people. That matters for the law.
The new laws, and the gap in them
States have started to act, though not all in the same way.
| State | What changed |
| New York | Since November 2025, a business must disclose when an algorithm set a price using your personal data. The required wording is close to: "This price was set by an algorithm using your personal data." |
| Connecticut, Maryland | Laws restricting surveillance pricing took effect on 1 October 2026, in Maryland aimed at larger food retailers and delivery platforms |
| New Jersey | A ban on surveillance pricing for groceries was signed in July 2026 and is due to take effect about a year later |
| Federal | The FTC is studying personalized pricing on delivery platforms. More than 40 bills are pending across two dozen states |
Notice what most of these laws target: prices built from your personal data. A price set for a whole restaurant from its sales history and its customers' estimated tolerance may not fall under them at all. So the most common form of AI pricing you meet, the store-by-store kind, is likely to stay legal for a while. Your best protection is still your own habits.
Five ways to stop overpaying
1. Compare stores in the app before you drive. The McDonald's app shows the price of the restaurant you have selected. Switch the store and the price can change. Thirty seconds of checking tells you which nearby location is the cheap one, and you only need to find out once.
2. Avoid the captive locations. Highway service areas, turnpikes, airports, stadiums and tourist strips are where customers have the fewest alternatives and push back least. In the engine's own vocabulary, low sensitivity. Eat before, or after.
3. Use the deals, not the menu board. Chains push app offers and value bundles precisely because they let the company discount for price-conscious customers while keeping menu prices high for everyone else. If you do not use them, you are the customer paying full price.
4. Treat price as a vote. The engine learns from what you do. A store where customers keep buying after each increase is graded as one that can take more. Going two miles further, or skipping a visit, is the only signal it understands.
5. Watch the same pattern elsewhere. Groceries, delivery fees, tickets, rides. If two places or two apps quote different prices for the same thing, compare before you pay. Where you live in New York, look for the algorithmic pricing disclosure.
Who wins, and the honest other side
It is fair to say what McDonald's would say. Costs really do differ between locations: rent, wages, local taxes and traffic. Pricing every restaurant the same would mean losing money in some places or overcharging in others. Franchise owners are small business owners with their own margins, and many will see a tool that helps them price well as useful, which is the same logic small businesses use across the AI math for owners.
The problem is the input the system is said to weigh most. Charging more because rent is higher is a cost. Charging more because customers at that store are less likely to walk away is a different thing. It turns convenience, habit and lack of choice into margin. Most customers never see it, because they only ever see one menu board at a time.
What this means for you
If you eat fast food regularly, spend five minutes once comparing the restaurants near you in the app. For a family that eats out weekly, that single check can be worth a couple of hundred dollars a year.
If you travel or commute by car, assume that the most convenient location is the most expensive one, and plan your stops.
If you run a small business, pricing tools like this are becoming cheap and common. Pricing to cost and to local demand is normal. Pricing purely to how trapped your customers feel tends to cost you their trust the day they compare.
The honest take
What I find telling about this story is how ordinary it is. Nobody stole data. Nobody broke a law. A company used its sales history and some clever statistics to work out where customers would quietly accept a higher price, and charged it there. That is what AI pricing mostly looks like in practice: not a dramatic invasion of privacy, but thousands of small, invisible decisions about how much you will tolerate.
The good news is that these systems learn from behaviour, and behaviour is the one thing every customer controls. The engine thinks your store can take more because customers there keep paying. The cheapest fix in the whole story is to stop being the customer it has you down as.
So, before your next order: do you know what the McDonald's two miles from yours charges?
Frequently asked questions
Why are some McDonald's more expensive than others?
Franchise owners set their own prices, and costs such as rent and wages differ by location. Reuters also reported that McDonald's uses a machine-learning engine that recommends prices for each item at each restaurant, partly based on how much that store's customers are estimated to be willing to pay.
Does McDonald's use AI to set prices?
According to a Reuters investigation published in September 2026, McDonald's has used a machine-learning pricing engine across nearly 14,000 US restaurants since at least 2019. McDonald's describes it as a tool that gives restaurant-specific recommendations, with franchisees keeping the final decision.
How big can the price difference be?
Reuters found two company-run McDonald's in Fresno, about two miles apart, charging $5.69 and $6.89 for a Big Mac, a 21% gap. It could not confirm that the pricing engine alone caused it.
How much have McDonald's prices gone up since 2019?
By McDonald's own 2024 figures, the average Big Mac rose from $4.39 to $5.29, up 21%, and the average menu item rose about 40%. Independent surveys have claimed larger rises.
Is AI pricing legal?
Generally yes. New York requires disclosure when an algorithm sets a price using your personal data, and Connecticut and Maryland restricted surveillance pricing from 1 October 2026. Most of these laws focus on prices built from personal data, so store-by-store pricing may not be covered.
How can I avoid paying more?
Compare nearby restaurants in the app before ordering, avoid highway and airport locations, use app deals and value bundles, and compare prices across stores and apps for groceries and delivery too.
Sources: Yahoo Finance on the Reuters investigation into McDonald's pricing tool; Decrypt: McDonald's AI pricing engine gauges what customers will pay; Delana: the 21% Big Mac gap and the surveillance-pricing laws; McDonald's: Providing meaningful value to our fans, with a side of facts; ConsumerAffairs on the Instacart price study; Yahoo Finance: Instacart ends its item price tests.
