How to demand forecast in an ‘appetite economy’

PwC/Numerator data show GLP-1 adoption reaching roughly 1 in 5 U.S. households by late 2025, up from less than 1 in 10 a year earlier. In households where a GLP-1 user is the primary grocery shopper, food spending falls 4%-6% in the first year, with single-person households down 7%-9%; QSR spending drops about 5% and casual dining about 2%. The article points to a meaningful demand shift for food, beverage, and restaurant operators as consumers reallocate toward fresher, protein-forward, nutrient-dense products.

Analysis

The key second-order effect is not just lower calorie intake; it is a re-bucketing of household baskets toward higher-margin, higher-turn products. That is structurally better for grocers, club channels, and health-forward private label than for center-store snacking, frozen indulgence, and value QSR traffic, because the consumer still spends but with a different mix and lower trip frequency. The 4%-6% grocery spend decline in GLP-1 households also implies slower unit growth for packaged food even when top-line dollars hold up via inflation.

The biggest near-term winner is the retail ecosystem that can capture “better-for-you” substitution without losing the basket to specialty formats. Protein, yogurt, seafood, and produce suppliers should outperform broadline branded CPG because GLP-1 users appear to trade down on quantity but trade up on nutritional density; that usually compresses branded share and increases private label penetration in staples. The loser set is more cyclical than structural in the first 6-12 months, but if adoption keeps rising, restaurant traffic pressure becomes a margin problem rather than a same-store-sales problem as operators lean on discounting to defend frequency.

The market is probably underestimating how long this takes to show up in earnings. The current data suggest a gradual demand leak, not a cliff, so consensus may be too aggressive in dismissing the effect as a one-time mix shift; the more important risk is compounding over 2-3 years as oral access and insurance broaden adoption into mainstream households. A reversal would require either meaningful affordability setbacks, intolerable side effects, or a faster-than-expected rebound in lower-cost obesity alternatives that do not alter eating behavior as much.

The contrarian angle is that the bearish read on food is too broad. If total spend only falls modestly, the profit pool migrates toward companies with pricing power in premium protein and away from commodity snack exposure; that makes this more of a relative-value rotation than an outright demand collapse. In other words, the winner is not ‘less food’ but ‘different food,’ and the first derivative of margin improvement may show up before the volume headwinds become obvious in reported sales.

Key Decisions for Investors

• Long WMT / COST vs short CPB or GIS over the next 6-12 months: retail mix should benefit from premium protein, fresh, and private label substitution while packaged snack/center-store names face persistent unit pressure; target 10-15% relative outperformance if GLP-1 penetration keeps compounding.
• Long MDLZ put spread or short MDLZ into earnings over 1-2 quarters: exposure to discretionary snacking makes this one of the cleaner ways to express slower impulse purchase frequency; risk/reward improves if management comments on volume elasticity worsen.
• Pair trade long TSN / short a restaurant basket (e.g., CMG, MCD) for 3-6 months: protein demand should be relatively resilient while QSR traffic faces frequency pressure; look for 300-500 bps spread if share shift toward at-home protein accelerates.
• Long GIS/CPB hedge with XLP call spreads only if valuation cheapens further: avoid outright shorting staples indiscriminately because the basket is rotating, not vanishing; use options to cap downside in case GLP-1 adoption slows.
• Buy OTM puts on YUM or SBUX around the next two reporting cycles if same-store traffic is already soft: these names have the most leverage to fewer visits per household, and the thesis can play out over months rather than days.