New York Attorney General Letitia James has sent a document demand to Instacart seeking detailed information about the company’s pricing practices and “price-setting experiments,” citing concerns that Instacart’s disclosures may not comply with New York’s Algorithmic Pricing Disclosure Act (G.B.L. § 349-A). The Attorney General’s office says Instacart must clearly and conspicuously disclose when consumer personal data is used to affect prices—and that the disclosure must appear near prices.

In a Jan. 8, 2026 press release announcing the letter, the Attorney General’s office pointed to a recent Groundwork Collaborative and Consumer Reports study alleging Instacart users were shown materially different prices for the same items, including instances where some shoppers saw prices “up to 23 percent higher” for the same products in the same store at the same time.

What’s new

• The AG’s office has opened an enforcement-style inquiry via a formal letter demanding documents and data about Instacart’s pricing tools and tests, with a Jan. 29, 2026 production deadline.

• The AG explicitly argues Instacart’s current disclosure approach—presented via fine-print links and not shown on category pages or product pages—may fail the statute’s “clear and conspicuous” standard and the requirement that disclosure appear for all covered price displays.

What happened

In the Jan. 8 letter to Instacart CEO Chris Rogers and General Counsel Morgan Fong, the Attorney General requests information about reports of substantial price variation on Instacart and Instacart’s compliance efforts with G.B.L. § 349-A, which the letter states took effect Nov. 10, 2025.

The law defines “personalized algorithmic pricing” as dynamic pricing set by an algorithm that uses personal data, and defines “personal data” broadly as data linked directly or indirectly to a specific consumer or device.

The statute also specifies required disclosure text. The letter quotes the requirement that covered entities include, with any displayed/advertised covered price, a clear and conspicuous disclosure stating: “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA.”

What Instacart must produce by Jan. 29

The AG’s document requests are operational and granular, including:

• Retail partner agreements with named retailers, including provisions on prices, markups/discounts, loyalty integrations, and use of tools like Eversight and Caper Cart—plus prior versions if changed around Instacart’s Dec. 22 announcement about stopping certain tests.

• CPG brand agreements with named brands involving pricing/discount terms and the use of Eversight or Caper Cart—again including prior versions if changed.

• Information about pricing/promotion tools referenced in the letter (including Eversight and Caper Cart), and how they are used in connection with pricing and promotions on Instacart, including whether personal data factors into any price-setting decisions.

• A detailed description of all price experiments, including dates, selection/assignment logic, variables adjusted (e.g., sale price, original price, discount/promotion), whether customer data is used, and the effect on totals paid.

• Screenshots and screen flows showing how price/discount/promo information and disclosures appear across web and mobile from the first price display through the transaction.

The letter instructs Instacart to send responses to the Assistant Attorney General by January 29, 2026.

Why this matters for subscription operators

This isn’t a grocery-only story. It’s a clear signal that pricing logic and experimentation are becoming enforcement-grade surfaces—especially where personalization is involved.

• First, pricing A/B tests are now part of the compliance conversation when personal data is in the loop. The letter specifically references “randomized tests” and the use of personal information (such as delivery address) and “personalized incentives” in explaining why disclosure is being triggered. For subscription teams, that maps directly to segmented intro offers, cancellation save offers, winback offers, and other price/discount tests.
• Second, “personal data” is defined broadly enough that “it’s just an experiment” may not be a safe framing. New York’s law includes data linkable to a consumer or device, which is exactly what most testing frameworks use to assign and persist variants. The practical takeaway: if your pricing tests rely on identifiers, behavioral signals, or targeting logic, assume regulators may ask you to explain what data drives which price.
• Third, disclosure implementation risk is product risk. The AG argues that disclosures buried behind fine-print links and not shown near prices may not satisfy the “clear and conspicuous” requirement. For subscription businesses, this is the same problem pattern as cancellation and renewal disclosure enforcement: it’s less about what your policy says and more about what the user sees at the decision point.

This is also one of those moments where it’s worth bringing legal (and privacy) expertise in early—not to shut down price testing, but to sanity-check where personalization can trigger disclosure requirements and to make sure your pricing, offer, and cancellation flows would hold up if a regulator asked to see them.

INSIDER TAKE

If you run subscriptions, the real signal here is that regulators are starting to treat “pricing systems” the way they treat cancellation flows: as something that can be tested, documented, and challenged at the UX layer. New York’s statute isn’t asking for a policy page; it dictates a specific disclosure and expects it to be shown near prices when covered pricing is used.

The other thing to notice is the level of detail the AG wants. The letter asks for the mechanics of experimentation—who was selected, what variables changed, whether customer data was used, and what the impact was on what people actually paid. Even if your team views pricing tests as routine growth ops, this is a reminder that those artifacts are discoverable, and they need to be legible to someone outside the building. Not because you’re doing something wrong, but because you may need to demonstrate you’re being deliberate, consistent, and transparent.

Finally, the partner angle is not a side plot. The AG is asking for agreements with retailers and CPG brands and how tooling influences promotions and discounts. Subscription operators live the same reality with app stores, channel bundles, affiliates, and payment partners: customers still experience the outcome as “your price.” The most practical move is to align product, legal, and growth on a plain-English description of what influences price (and what doesn’t), where it’s disclosed, and how you document tests—before you’re forced to do it under a legal deadline.