Maryland Just Banned What Walmart Is Building: Inside Grocery's AI-Pricing War

On April 28, 2026, Governor Wes Moore signed the Protection From Predatory Pricing Act (HB 895) into law, making Maryland the first state to prohibit food retailers and third-party delivery platforms from using personal data to set individual grocery prices. The statute takes effect on October 1, 2026 — six weeks before the U.S. holiday grocery season and within days of Walmart's stated deadline to put digital shelf labels (DSLs) on every shelf in all ~4,600 U.S. stores. Penalties reach $10,000 per violation and $25,000 for repeat violators, enforced by the Maryland Attorney General's Consumer Protection Division.

On August 19, 2026, the Federal Trade Commission released a proposed Enforcement Policy Statement on Personalized Pricing — its first formal guidance tying Section 5 of the FTC Act to AI-priced milk, bread and other staples. The proposed policy uses a grocery scenario as its worked example: a chain charging a delivery customer a higher price for milk based on data showing children live in the household. The public comment period was extended to September 18, 2026, with the FTC signaling its intent to act against any pricing practice that uses personal data without clear, conspicuous disclosure.

Disclosure requirements matter less for retailers than the political optics. A 2025 Consumer Reports survey of 2,240 U.S. adults found 72% of recent Instacart users did not want AI to charge different users different prices for any reason — a number that explains why bipartisan bills from Senators Merkley (D-Ore.) and Luján (D-N.M.) are already filed in Congress to extend the Maryland model nationwide.

On December 9, 2025, Consumer Reports and the Groundwork Collaborative published the results of 400 volunteer online-shopping sessions in September 2025. About three-quarters of products checked showed different prices for different customers on [Instacart] (Maplebear). Per-item variance ranged from 7 cents to $2.56, and a single Safeway Seattle test on September 4 — 39 volunteers, 20 products — produced five different basket totals spanning $114.34 to $123.93, a swing that Consumer Reports estimated as roughly $1,200 in extra annual grocery spend for a typical four-person household. Experiments were observed at Albertsons, Costco, Kroger, Safeway, Sprouts Farmers Market, and Target.

The lesson for retailers: a 2–5% margin lift from personalized pricing is large relative to [Kroger]'s 1.28% or [Albertsons]'s 0.86% operating margin — but it is now politically indefensible in any state that follows Maryland's lead. Pricing tools persist, but the personalization layer has been amputated.

[Walmart] began the VusionGroup DSL pilot in 2024 and announced in early 2026 that all 4,600 U.S. stores would carry the digital tags by year-end — an acceleration from the prior 2,300-store target. Per the company, the rollout cuts pricing-task time by 75% and reduces pricing errors by 98%. On September 27, 2026, Walmart publicly denied it uses personal information to set prices, framing the DSL deployment as a labor-efficiency story rather than a surveillance-pricing one.

Strip the personalization story away and the deployment still changes the math. A 75% reduction in shelf-pricing labor across 4,600 stores is meaningful for a retailer whose consolidated [Walmart] FY2026 (ended Jan 2026) operating expenses were $147.9B — a labor-cost line where single-digit-percentage gains show up clean on the income statement. [Kroger] confirmed in early 2026 it is using a Microsoft Azure pricing engine layered with Intelligence Node's competitive-pricing algorithms and is running Google Cloud's Gemini agentic platform for personalization that complies with whatever the FTC and state bans permit.

When grocers gain real-time elasticity data on every SKU, supplier pricing power erodes — and the supplier income statements now show the trough. [Kraft Heinz] reported a $4.7B operating loss for FY2025 on February 11, 2026, of which $9.3B was non-cash goodwill and intangible impairment. [Conagra Brands] posted a (20.1)% reported operating margin for Q2 FY2026 (reported December 19, 2025), with non-cash goodwill and brand impairment charges producing a $1.39 reported diluted loss per share versus an adjusted $0.45 — adjusted operating margin was still only 11.3% on a 6.8% revenue decline. [General Mills] reported on July 1, 2026 that fiscal 2026 restructuring, transformation, impairment, and other exit costs totaled $3.0B of net expense.

[Costco]'s FY2026 (ended Aug 2026) operating margin was 3.85% on $303.2B of revenue — lower than [Kroger] in absolute terms — but Costco runs ~87% of revenue through cost of goods, an unusual structure that reflects the Kirkland Signature membership-economics flywheel. The Consumer Reports study named Costco among the retailers whose customers saw varied pricing on Instacart, but Costco's bulk-SKU cadence, its membership gate, and its leadership in keeping prices identical store-to-store give it the lowest regulatory exposure to the Maryland-FTC agenda.

[Albertsons] sits at the other extreme. With a 0.86% FY2025 operating margin on $83.2B of revenue and net income of just $217M, the company has the most to gain from any pricing optimization and the most to lose from any ban that constrains it. Maryland is a smaller share of its footprint than for [Kroger], but every state-level replication of the Maryland template eats into the case for [Albertsons]'s pricing-tech capex.

In the next 6–18 months, the dominant catalysts are regulatory. Watch for (1) additional state-level replications of Maryland HB 895 (the bipartisan federal bill from Sens. Merkley and Luján is the escalation path), (2) the FTC finalizing its August 19 proposed policy statement with Section 5 enforcement actions, and (3) any successor to [Instacart]'s Eversight — the most likely vendor replacement is Revionics, Blue Yonder, or Intelligence Node, all of which already serve [Kroger] and other major grocers and are now aggressively marketing \"personalization-compliant\" variants. The Maryland ban's effective date on October 1, 2026 is the next hard date.

Over the 1–3 year horizon, the deeper story is the supplier-margin transfer. Retailers' DSL-driven labor savings (75% task reduction at [Walmart]) and elasticity-driven SKU-level trade-promotion optimization compress supplier negotiating room; the [Kraft Heinz]/[Conagra Brands]/[General Mills] impairments are the leading evidence. If [Instacart] (TTM revenue $3.99B, gross margin 72.4%) decides to refocus on ad revenue rather than re-launch Eversight, the public-perception overhang eases but the margin-pool shift to retail — and away from CPG — does not reverse.