The calorie cut arrives twice

GLP-1 receptor agonists slow gastric emptying and amplify satiety. People eat fewer occasions and smaller portions. They also, in the survey work, change what the remaining occasions are. Protein and fibre stay. Ultra-processed density falls. Sweet and fried impulse packs take the first cut. Alcohol is in the same bucket. That is not a branding exercise. It is a stomach.

Fewer units

Walmart US CEO John Furner said the quiet part in public: a slight pullback in the overall basket, fewer units, slightly fewer calories, among customers on the drugs. When the largest grocer in the country can see it in the till, manufacturers do not get to call it a lifestyle column. Numerator's quarterly tracker, 95,000-plus responses, now has a current GLP-1 user in 22% of households, double October 2023. Those households spend about 4% less at the grocer than similar households not on the drugs, after a year. They account for more than $660 billion of spending across CPG, general merchandise, and QSR.

Attest's May survey of 1,000 adults aged 18 to 67 found that 14% are currently taking the drugs and a further 9% have used them, 23% lifetime, higher than Gallup on a smaller and younger sample. Among current users, half are eating smaller portions and nearly a quarter are buying less food overall. Sweet snacks and alcohol are the first things to go in the self-report, 46% and 45% cut back. Millennials, not boomers, are the heaviest current users at 18%, then Gen X at 17%, boomers at 8%. That age mix is the opposite of the 2022 Medicare-and-diabetes picture. It is a grocery picture.

OC&C's 2026 cut puts the current-usage volume trim on US food and beverage at roughly 1 to 2%. That sounds small until you remember packaged food is a low-single-digit volume business in a good year. EY-Parthenon has the decade-scale number: as much as $12 billion off US snack sales if the diet shift sticks. Circana, meanwhile, still prints a growing snack market, $231 billion in the US, +3.2%, because price, occasion, and functional packs are doing the work that units are not. Fifty-five percent of US shoppers now eat three or more snacks a day, up nine points since 2021. Frequency and calories are not the same variable.

A different basket

The same surveys that show fewer units show a rewrite of the cart. Numerator: away from pantry carbs and bakery, toward protein, fibre, seafood, functional nutrition, and more intentional snacks. Circana's 2026 Global Snack Unwrap: users shift toward nutrient-dense foods, protein, nuts, grains, and pull back on chocolate, sweet, and fried. Attest: 60% of Americans increased fresh fruit and vegetable purchases in the last six months; 51% bought more high-protein foods. Household spillover, the non-user copying the user's plate, is how an 11% adult share becomes a larger aisle event. Room-mates, partners, and parents do not keep two pantries.

Nestlé built Vital Pursuit as a GLP-1 brand. A Nestlé USA executive told the trade press that only about a fifth of those buyers take a weight-loss drug; the Associated Press put the same fact the other way: most of the sales sit in households where nobody is on the medicine. Four in five people buying the GLP-1 meal are not on GLP-1. The badge recruited a shopper who wanted permission to pay more per calorie. That is why mix can print as growth while mouths leave the large bag.

Foodservice splits the same way. Classic QSR bundles and impulse delivery take the volume hit. Full-service and high-protein builds take some of the mix. The LSEG count Reuters used in February is the corporate tell: nearly three dozen non-healthcare companies mentioned GLP-1s or weight loss on earnings calls in the first stretch of 2026, up from 14 a year earlier and five the year before that. The mentions are not insight. The volume/mix line in the same filing is.

Private label is the other mix shift, and it is easy to blame on GLP-1 when it is also a price cycle. Circana has private label at 16.6% of US snack unit share. Europe is 46.6%. The US number can rise because a GLP-1 household wants a smaller, cheaper pack of something they still eat, or because a non-user household is tired of paying 2022 to 2024 list. Both print as volume/mix. Only one is the drug.

What the companies already said

PepsiCo is the scale case. Foods are about 58% of the company. PepsiCo Foods North America net revenue fell 2% in the latest quarter; first-half organic sales were down 0.5%. International organic revenue rose 7% over the same stretch, Europe, Middle East and Africa +6%, Latin America +4%, Asia Pacific Foods +12%. Volume in the North American food business was flat in the quarter ended 13 June even after price cuts of up to 15% on Lay's, Doritos, Cheetos, and Tostitos. Volumes have fallen in four of the last six quarters. Ramon Laguarta called it a subdued consumer environment in February and talked about price adjustments and a simpler portfolio. PwC's read of Numerator had GLP-1 adoption at 21% of US households in May 2026, from 9% in January 2025. PepsiCo's international book is the hedge: GLP-1 use is about 11 to 12% of US adults, closer to 7% in the UK, lower still on the continent, and cost-constrained in Mexico and Brazil. The calorie hole is, for now, an American profit and loss.

Hershey has named rising GLP-1 use as a reason US demand for indulgent, impulse-driven snacking is changing, and it is leaning on international expansion, salty snacks, and better-for-you formats to offset the core at home. Mondelez and Kraft Heinz have printed some version of the same US-soft, overseas-less-soft split. Campbell's is closing the Hyannis, Massachusetts potato-chip plant as part of a snacks-network optimisation.

Hostess is the exposed aisle

J.M. The Hostess deal was $5.6 billion in 2023. Fiscal 2026's first quarter Sweet Baked Snacks net sales, excluding the Voortman, Big Texas, and Cloverhill divestitures, were down 10%, with volume/mix off 8 points and price off 2, mainly in snack cakes. On the August 2025 call Mark Smucker said they still saw no meaningful impact from GLP-1s in their categories and would keep offering reduced-sugar and smaller-portion variants. Those two sentences can both be true. Hostess can be a distribution and convenience-store problem while GLP-1 rewrites the category around it. Coffee can carry a quarter while the cake aisle shrinks. Uncrustables can carry the volume story in a quarter snack cakes are not. The useful line in any Smucker filing is US Retail volume/mix, and whether the year is still being guided as price and mix rather than more units.

The leak is discontinuation

It is that nearly two-thirds of former users discontinued inside six months. The market is still largely shaped by recent adopters. If persistence stays this bad, the volume story is a treadmill: new scripts have to replace quitters before they can shrink the aisle. Gallup's compounded-vs-brand split is the commercial version of the same leak. Compounded users are much more likely to have switched from brand than the other way around, 35% vs 10%, and two-thirds of those switchers name cost or insurance.

Pharma-adjacent social feeds talk scripts, shortages, and orals. Food feeds talk Hostess plants and Frito-Lay volume. They are looking at the same drug and scoring different profit and loss statements. If persistence improves, coverage, cheaper orals, fewer GI drop-offs, OC&C's 1 to 2% trim is the floor, not the ceiling. If persistence stays ugly, the aisle loses a rotating 11% and gets most of them back, and the structural story is mix, protein, fibre, smaller packs, not extinction.

What volume/mix actually is

In a US food 10-Q, net sales are usually bridged in two or three lines: price realization, volume/mix, and sometimes acquisitions or divestitures. Volume/mix is itself a blend. More units of the same pack is volume. A shift from a large bag to a small bag, or from a value cake to a protein bar, is mix. A company can print volume/mix of minus 4% because people bought fewer bags, or because they bought the same number of smaller bags, or because the heavy SKU lost and the expensive SKU won. The GLP-1 story produces all three at once. That is why a single volume/mix print is a test and not a conclusion, and why the guide on volume versus price is the second half of the read.

Oral GLP-1s and Medicare coverage are the 2027 and 2028 risk to the discontinuation leak. Scripts are a flow. The grocery profit and loss is a stock of people who are still injecting, still swallowing, and still buying dinner for a household that has started to copy the plate. Novo Nordisk and Eli Lilly sit on the other side of every line. Prescriptions times the price they actually collect is the one number in this chain that is growing for sure.

A calorie is a calorie, and beer already had demographics against it. Do not let a drug label do all the work on a category that was shrinking first. Sugar and corn will move with the Brazilian crop. They belong on a commodity strip. They do not tell you what happened to Hostess.

Staples leading a down week for the index is the short truth. It will be quoted as a refutation of a five-year volume curve. A refutation would look like snack volumes up for two quarters while prescriptions still grow fast, or prescriptions stall while volumes fall and margins expand and the stocks rise anyway. A Friday sector win is none of those.

Read units in the filing. Ignore the badge on the box.