According to CNBC, a study released Tuesday by Groundwork Collaborative, Consumer Reports, and More Perfect Union found that Instacart’s AI pricing tools are driving up costs for some shoppers. The research, which enlisted 437 shoppers across four cities, discovered that almost 75% of the items tested were offered at different prices within the same store on the Instacart app. In one stark example at a Washington, D.C. Safeway, a carton of Lucerne eggs was listed at five different price points. The total cost for an identical basket of goods at a single store varied by about 7%, which Groundwork calculates could mean a “cost swing of about $1,200 per year” for a regular user. The experiment was run at major partner retailers like Target, Costco, and Kroger.
The Business of Personalized Prices
Here’s the thing: this isn’t just a glitch. It’s the business model. Instacart, and platforms like it, use AI to analyze a staggering amount of data—your location, your ordering history, maybe even how long you linger on an item page—to guess exactly what you’re willing to pay. They call it “dynamic” or “personalized” pricing. But let’s be real, it’s personalized profit-maximizing. The immediate beneficiary is, of course, Instacart and its retail partners, who can squeeze more margin out of customers who are less price-sensitive or in a hurry. It’s surge pricing for eggs and milk. And because you can’t easily compare your Instacart cart price with your neighbor’s, the lack of transparency is a feature, not a bug.
Why This Feels Different
We’re used to airline tickets and ride-share fares changing by the minute. But groceries? That feels like a violation of a basic trust. A gallon of milk at Store X should cost what it costs. This study shows we’ve quietly crossed a line where the price on the shelf (or the app) is no longer a fact, but a suggestion tailored to you. So what’s the trigger? Is it because you order organic every time? Live in a wealthier ZIP code? The algorithm won’t tell you. This murkiness is where the real power—and risk—lies. It turns every shopping trip into a silent, one-sided negotiation where you don’t even know the rules.
The Bigger Picture on AI Transparency
Look, algorithmic pricing is everywhere now. But this study is a concrete, relatable example of why we need to demand more transparency. It’s one thing for an industrial supplier to use complex models for industrial panel PCs in a B2B setting. It’s another to apply the same opaque logic to a consumer’s weekly food budget. When the cost of feeding your family becomes a variable based on secret signals, it erodes trust in a fundamental way. The question isn’t just whether this is legal (it probably is), but whether it’s right. And as AI gets better at profiling us, this kind of personalized pricing is only going to get more precise, more pervasive, and harder to spot.
