Some grocers are using AI to cut food waste and boost profit margins

Grocers are using AI-driven dynamic pricing to reduce shrink and capture value from perishable inventory, with Flashfood saying partners have cut shrink by an average of 27% while adding incremental traffic. The model aims to protect margins without blanket markdowns, and Kroger is expanding Flashfood to more than 100 additional stores this month. Analyst Bill Kirk highlighted Kroger’s data advantage and maintains a $78 price target versus a Thursday close of $67.77.

Analysis

The key equity implication is that this is less a grocery-margin story than a data-monetization story. Chains that can translate perishables optimization into a closed-loop pricing engine should see a structural advantage in gross margin volatility, working capital efficiency, and customer retention, while smaller grocers without scale data likely face a widening productivity gap. That favors KR as the clearest public beneficiary, with the second-order effect that its private-label and fresh categories can be defended without resorting to broad-based price cuts.

The market may be underestimating how fast this behavior scales across the basket. Once shoppers learn that one retailer systematically offers better near-expiry value, trips become more fragmented and price transparency rises, which pressures everyone else to either match on precision or bleed shrink. DG is the most exposed on the negative side because value-seeking traffic can become more promotional and less sticky if grocery baskets migrate toward more targeted markdown ecosystems; COST is comparatively insulated because its membership model and bulk value proposition already capture the same budget-conscious consumer without requiring per-item discounting.

The main risk is execution, not concept: if AI-led markdowns are applied too aggressively, retailers can train customers to wait for discounts and compress fresh-category gross margins rather than improve them. The payoff is likely months, not days, because the near-term earnings contribution is modest while the real benefit comes from lower shrink, better inventory turns, and improved customer data over several quarters. A reversal would come from deflation in food inflation or a normalization of consumer budgets, which reduces the urgency for value-hunting and lowers the economic value of dynamic pricing.

Contrarian view: the consensus is focused on shrink reduction, but the bigger opportunity is basket expansion from incremental trips and cross-sell. If the platform consistently drives customers into stores for discounted perishables, the full-price attach rate can matter more than the markdown itself, which creates a more durable earnings lever than one-off waste reduction. That makes this a quality-of-data winner rather than a simple cost-cutting story.

Key Decisions for Investors

• Long KR vs short DG for 3-6 months: KR should monetize its data stack into margin and traffic gains, while DG is more vulnerable to value-shopping fragmentation; target 1.5-2.0x relative outperformance if the thesis plays out.
• Add KR on pullbacks ahead of the next earnings cycle: the market is likely underpricing a multi-quarter gross margin lift from lower shrink and better fresh pricing; use a 6-12 month horizon and trim if the stock rerates to reflect most of the productivity gain.
• Avoid chasing COST on this theme: the upside from AI markdown optimization is real but incremental; prefer it as a defensive long only if paired against a more exposed grocer or retailer with weaker data capabilities.
• Use DG as a tactical short or underweight if consumer value-seeking intensifies over the next 1-2 quarters: downside comes from continued trip fragmentation and promotional pressure, though the trade should be tightly risk-managed if broad discount retail sentiment improves.
• Optionality idea: buy medium-dated KR calls into any weakness if management reiterates AI-driven productivity gains; the asymmetry is better than outright stock ownership if adoption headlines accelerate over the next 2-4 earnings prints.