Have you ever added the exact same loaf of bread to your online cart, hit checkout with a few hours later, and suddenly it costs two bucks more?
I have. And until last week I just chalked it up to “prices change, whatever.” But a brand-new investigation dropped on Tuesday that made me feel a lot less crazy — and a lot more annoyed.
Turns out, one of America’s biggest grocery delivery services has been quietly running AI-powered pricing experiments that make identical shopping carts cost noticeably different amounts for different people — sometimes from the very same store, on the very same day.
And when I say “noticeably different,” I’m not talking about a few cents. We’re talking swings big enough to add more than a thousand dollars to a family’s yearly grocery bill.
The Study That Changed Everything
A coalition of consumer advocacy groups decided to put the rumors to the test. They recruited hundreds of real shoppers in four different cities and had every single one of them build the exact same shopping list from the exact same stores using the same popular delivery app.
The results were jaw-dropping.
Nearly three-quarters of the items showed up at different prices depending on whose phone was looking at them. In one particularly wild example, a single carton of a dozen basic eggs appeared at five separate price points when shoppers checked a single supermarket location in Washington, D.C.
The total basket price varied by roughly 7% on average. Doesn’t sound huge, right? Until you do the math: for the typical American household spending about $6,000–$7,000 a year on groceries, that’s an extra $400–$1,200 disappearing from your wallet without you ever realizing why.
“These aren’t sales or promotions. These are the same products, same store, same moment in time — yet the app is quoting wildly different prices to different customers.”
– Lead researcher from the study
How Does This Even Happen?
Here’s where it gets technical — but stick with me, because it matters.
A few years ago, the delivery company in question bought a lesser-known startup that specializes in “AI-powered pricing and promotions.” Think of it as the same kind of algorithm airlines and ride-sharing apps use to decide whether your flight or Uber should cost $79 or $279.
Only now it’s being pointed at your milk, eggs, and toilet paper.
The official line from the company is that these tests are “limited,” “short-term,” and only look at “category-level price sensitivity.” Translation: they claim they’re not spying on you specifically — just trying to figure out how much the average person will tolerate paying for cereal before they switch brands.
They also insist prices never move in real time based on demand (no surge-pricing your bananas at 6 p.m. on Sunday) and that zero personal data is involved.
But here’s what raises my eyebrow: if nothing shady is happening, why do identical carts built minutes apart show completely different totals? And why does the company refuse to let customers opt out of being guinea pigs?
This Isn’t Just One Company’s Problem
Let’s be real — this isn’t isolated. We’ve all heard stories about concert tickets doubling the moment you log in from a MacBook instead of an older Android. Or hotel rooms jumping $80 a night when you clear your cookies (or don’t).
What makes groceries different is how essential they are. Food isn’t a Taylor Swift ticket you can decide to skip. When algorithms start playing games with the price of baby formula or bread, it stops feeling like clever business and starts feeling… predatory.
- Airlines have been doing dynamic pricing forever
- Ride-sharing invented surge pricing
- Hotels and e-commerce sites A/B test prices constantly
- But groceries? That hits different.
There’s something particularly gross about using AI to extract the absolute maximum from people who are literally just trying to feed their kids.
The Bigger Trend: Surveillance Pricing Is Everywhere
Regulators are finally waking up. The term you’re going to hear a lot more of is “surveillance pricing” — when companies use data about you (or people like you) to personalize how much they charge.
New York just passed the first state law requiring companies to disclose if they’re doing it. The FTC launched a formal inquiry last year. Even Congress is floating bills that would ban the practice when it relies on personal data.
Because here’s the quiet part they don’t say out loud: the goal of these algorithms isn’t to make things cheaper for most people. The goal is to figure out exactly how much each individual is willing to pay — and charge them one penny less than they’d walk away.
It’s not about competition driving prices down. It’s about technology letting companies capture every last dollar of “consumer surplus” — the gap between what something costs to produce and what you’d actually be willing to pay before giving up.
“Perfect price discrimination has always been the holy grail of economics. AI might have just made it possible.”
– Economics professor interviewed for the study
What Can You Actually Do About It?
Feeling helpless? Same. But there are a few practical moves that still work in 2025:
- Shop in incognito/private browsing mode (it actually still helps sometimes)
- Compare the same cart across multiple apps — prices can differ dramatically
- Use a VPN to appear in a different city (yes, location still affects pricing)
- Consider picking up your own groceries when the order is large — many stores often reserve the lowest prices for in-person shoppers
- Support legislation — the more noise consumers make, the faster change happens
Honestly? The nuclear option that always works: walk away. Every time enough people abandon an app because it sends a louder message than any tweet or complaint form.
My Take — Why This One Hits Home
I’ve been writing about personal finance and tech for years, and I still got caught off guard by how normalized this has become.
We went from “May I take your order?” to “Let me run 400 data points about you through a black-box algorithm and decide how poor you look today” without most of us even noticing the switch.
And maybe the most frustrating part? Companies know we’ll grumble for a week and then keep clicking “Place Order” because cooking dinner after a 10-hour workday is hard and convenience is addictive.
But every time we shrug and pay the mystery surcharge, we’re voting with our wallets for more of the same.
So yeah — are you okay with AI quietly deciding how much your groceries should cost? Or is this finally the line?
Because from where I’m sitting, once the algorithm starts playing games with eggs and milk, nothing feels off-limits anymore.
Let me know in the comments if you’ve noticed weird price jumps in your own grocery apps. The more stories we share, the harder it becomes for companies to pretend this isn’t happening.
