How GLP-1 Drugs Change the Feasibility Case for Restaurants, Franchises, Food Manufacturers and Gyms

One in eight American adults now takes a GLP-1 drug, and the households around them spend measurably less on snacks, soda, pizza and beer. Restaurant chains say they cannot see it in their numbers. Lenders have not put it in a single credit file. This report sets out what the 2026 evidence actually shows, where it conflicts, and how MMCG now applies it as a sensitivity in feasibility studies for the asset classes it touches.

Every few years a demand variable arrives that the standard pro forma does not know how to hold. Fuel prices did it to travel centers. Remote work did it to downtown lunch trade. GLP-1 receptor agonists, the class of drugs sold as Ozempic, Wegovy, Mounjaro, Zepbound and now as pills, are doing it to anything that sells calories.

The feasibility problem is not that the effect is large. It is that the effect is real, measurable at the household level, concentrated in specific formats, and completely absent from the way restaurant, franchise, food-manufacturing and fitness projects are underwritten. A 2026 limited-service restaurant study that projects traffic off a demographic ring and a competitive set, and never asks what share of the ring's households is on a GLP-1, is modeling a market that no longer exists in quite that form.

This report is written for the lender and the sponsor rather than the nutritionist. The medical content is one paragraph. The rest is demand, revenue and debt service.

Three years ago the question was whether these drugs would reach mass adoption. That question is settled. The KFF Health Tracking Poll found 12% of U.S. adults currently taking a GLP-1 for any reason in late 2025, double the 6% measured in May 2024, and the same 12% again in March 2026 (1)(2). Gallup, which asks specifically about weight loss, put current use at 11% in June 2026, up from 3% in 2024 and 8% in 2025 (3). On a U.S. adult population of roughly 265 million, those surveys imply 29 to 32 million current users.

Prescription claims show fewer. IQVIA counted about 11 million unique GLP-1 patients across all indications in the second quarter of 2025 (4). The gap between 11 million and 30 million is not a measurement error. It is the cash-pay, direct-to-consumer and compounded market that claims data cannot see. Gallup found 19% of current weight-loss users on a compounded product in mid-2026, and the manufacturers' own direct channels have moved millions of patients outside the insurance system entirely (3).

For a feasibility analyst the more useful unit is the household, because that is the unit that shops and eats out. Circana put households with at least one GLP-1 user at 23% of U.S. households in November 2025 (5). Numerator's tracker had 22% in August 2026, and PwC's May 2026 refresh had 21% (6)(7). Call it one household in five, and climbing.

What changed in 2025 and 2026 is price and access, and that is what makes the next leg of adoption different from the first. Novo Nordisk announced in February 2026 that list prices for Wegovy, Ozempic and Rybelsus will fall to $675 a month from January 2027, a cut of roughly half for Wegovy (8). Lilly's direct-to-consumer Zepbound vials run $299 to $449 a month (9). The Wegovy pill launched January 5, 2026 at $149 a month for the starter dose, and Lilly's Foundayo pill followed in April 2026 at the same entry price (10)(11). Medicare began covering obesity GLP-1s on July 1, 2026 at a flat $50 copay through the GLP-1 Bridge, now extended through 2027 (12). The access picture is not uniformly expanding: Medicaid obesity coverage contracted to 13 states in January 2026, and only 19% of large employers covered the drugs for weight loss in 2025 (13)(14). But the direction of the retail price is one way, and the pill removes the injection barrier that kept a large share of eligible patients out.

Two facts cut the other way, and they matter for how far the demand effect travels. First, persistence is poor. In a cohort of 125,000 U.S. adults with overweight or obesity and no diabetes, 64.8% had discontinued within a year and 84.4% within two years (15). Second, weight returns. A BMJ meta-analysis of 37 studies found patients regain about 4.8 kg in the year after stopping and return to baseline weight in roughly 1.7 years (16). So the population on the drug at any moment is a churning stock, not a permanent one, and the consumption effect for a given household is temporary unless the household restarts, which many do.

The forecasts have been chasing the data. Morgan Stanley raised its 2035 U.S. user projection to 55 million in April 2026, from 33 million, on the strength of Medicare access and the pills (17). J.P. Morgan projects 25 to 30 million Americans on treatment by 2030 (18). Both figures are already close to what the surveys say exists today, because the banks count branded, prescribed users and the surveys count anyone who says yes.

The single most important document for anyone underwriting a food business in 2026 is a Journal of Marketing Research paper published in December 2025 by Hristakeva, Liaukonyte and Feler. It links Numerator's purchase panel of about 150,000 households to repeated surveys of GLP-1 use, and it compares households that start a GLP-1 with matched households that do not (19). That design matters. Most of what circulates on this topic is a survey asking users whether they eat less, and users say yes. This paper measures receipts.

Within six months of adoption, households with a GLP-1 user cut grocery spending by 5.3%. Higher-income households cut 8.2%. Spending at limited-service restaurants, meaning fast food and coffee shops, fell 8.0%. Savory snacks fell 10.1%. Yogurt, fresh fruit, nutrition bars and meat snacks were the only categories to rise, and the increases were small against the declines (19)(20). The working-paper version of the same study, charted by McKinsey in June 2026, gives the category detail: chips and savory snacks down 11.5%, sweet bakery down 8.5%, frozen sides down 7.5%, cheese down 7.2%, cookies down 7.0%, soft drinks down 6.8%, candy and chocolate down 3.5%, and alcohol down just 1.4% (21).

Three later reads from the same panel family fill in the time profile. PwC's May 2026 refresh, covering about 110,000 households six to eight months after starting, found grocery down 5.5% and quick-service spending down 8.7%, with pizza down 22.2%, chicken down 11.6%, coffee and bakery down 8.0% and burgers down 6.4% (6). Numerator's quarterly tracker in August 2026 found the grocery gap narrowing to 3.8% after a full year, and reported that quick-service traffic among users "remains steady" while items per trip fell 3.7% (7). And Circana, which measures users against their own pre-adoption baseline rather than against a control group, found packaged food and beverage spending down only 1.6% after one year, with total restaurant spending actually up 0.9%, casual dining up 4.1% and quick service up 0.6%, alongside what it called steep declines in alcohol (22)(23).

Those figures do not agree, and the disagreement is informative. Matched-control studies capture the full effect of the drug relative to what the household would otherwise have done. Own-baseline studies let the effect wash out as the household's overall spending drifts with inflation and income. NielsenIQ's 2024 panel, which compared new weight-loss users to themselves a year earlier with no control group, found grocery spending up 22% in the first three months, because those households were also buying protein, produce and dairy as they changed how they ate (24). The honest summary is that the drug reduces a household's food spend by something between 2% and 6% depending on the window and the method, that the reduction is sharpest in the first six months, and that it falls heavily on a short list of categories: salty snacks, sweet bakery, soda, cookies, frozen convenience and pizza.

Surveys say the restaurant effect is far larger, and they are measuring something different. William Blair's April 2026 survey of 300 users found restaurant visits halved, from 9.8 to 4.9 a month, yet restaurant spending fell only 4%, which means the respondents who still go spend more per visit (25). The National Restaurant Association's own survey found GLP-1 users made 7.6 restaurant purchases in the prior week against 5.1 for non-users, which says more about who takes the drug than about what it does (26). Toast's August 2026 survey of 850 users found 51% eating out less, 28% more, and the cuts falling on desserts, appetizers, sides and drinks rather than on the visit itself (27).

Alcohol deserves its own line because the clinical and the commercial evidence point in different directions. The household panels show a small effect, 1% to 2% of alcohol spend. A randomized trial of low-dose semaglutide in 48 adults with alcohol use disorder found drinks per drinking day down about 30% by the second month against 2% on placebo (28). Both are true. The trial enrolled heavy drinkers; the panel measures household dollars including the non-users in the house and leaves out the bar tab. For a bar or a full-service restaurant the relevant number is somewhere between the two, and it is the one number in this report that no operator has been able to pin down.

Scaled to the national market, the Cornell effect sizes work out to roughly $9 billion a year in lost grocery spending and about $7 billion in lost limited-service restaurant spending (19). J.P. Morgan's February 2026 projection is an annual food and beverage revenue reduction of $30 billion to $55 billion by 2030 to 2034 (18). Against USDA's $2.58 trillion of total U.S. food spending in 2024, that is 1.2% to 2.1% (29). Morgan Stanley's calorie model lands in the same place from the other side: a 1.6% cut in annual U.S. calories by 2035, with the decline accelerating through 2028 and moderating after (17). Piper Sandler, revising upward in November 2025 for the pills and the price cuts, sees a 45 to 110 basis point annual headwind on packaged-food sales, cumulating to 325 basis points by 2028 (30).

A 1% to 2% national drag sounds like rounding. For a market that grows volume at 1% in a good year, it is the difference between growth and decline, and it does not land evenly. It lands on the formats and the categories named above, in the trade areas where adoption runs highest.

Here is the puzzle. The household data say GLP-1 households cut quick-service spending 8% to 9% within six months. One household in five is a GLP-1 household. Arithmetic says the quick-service sector should be carrying a drag of 1.5% to 2% against its counterfactual. And not one of the fifteen public restaurant companies whose disclosures were reviewed for this report says it can measure that in its own numbers.

McDonald's CEO Chris Kempczinski, February 11, 2026: "We don't yet see evidence of it really having a material impact on our business," with a caveat that "as adoption grows, we know that consumers' behavior changes" (31). Domino's CEO Russell Weiner, February 23, 2026: "We have not seen an impact on our business so far" (32). Brinker says it has seen the data everyone else has seen and has not experienced it. Darden is the only company in the set that has described an effect, and it has moved its view three times. In December 2024 it said the drugs "could be having an impact on the higher end brands" (33). In December 2025 CEO Rick Cardenas said "it's impacting drinking more than it's impacting eating, especially in our kind of brands" (34). By September 2026 he described GLP-1 usage as stable at about 12% of U.S. adults since mid-2025 (35).

What the companies have done, rather than said, is more telling. Restaurant Brands International and Dutch Bros both added GLP-1 drugs to the risk factors in their fiscal 2025 Form 10-K filings, RBI citing "increased adoption of weight loss medications such as GLP inhibitors" (36)(37). Chipotle launched a High Protein Menu on December 23, 2025 with two items labeled "GLP-1 friendly," and its CEO described a $3.80 High Protein Cup as "a solution to those looking for smaller portions, which is a fast-growing trend with the adoption of GLP-1s" (38). Starbucks put protein lattes and a $2 protein cold foam into every U.S. store in September 2025, and its CEO named GLP-1 users as one of the cohorts the foam targets (39). Sweetgreen's CEO told investors "as GLP-1 adoption increases, we will be a beneficiary," while reporting a 700 basis point drop in restaurant-level margin partly caused by larger protein portions (40). None of them has disclosed incremental same-store sales from any of it.

The industry traffic data explain why the effect is hard to see. Black Box Intelligence has chain same-store traffic negative in nearly every month from the start of 2025 through June 2026, running between minus 1% and minus 3% with a low of minus 5.7% in February 2025 (41). Black Box attributes that to inflation, a softening labor market and a gas price tipping point at $3.50, not to GLP-1. Circana's foodservice traffic for the full year 2025 was minus 0.3%, and its panel shows GLP-1 users cutting items per restaurant trip by 1% while visit frequency holds (42). The GLP-1 effect, in other words, is sitting inside a traffic decline three to five times its size that has a different cause, and it is showing up as mix (fewer sides, fewer drinks, smaller portions) rather than as fewer doors opened.

Sell-side analysts have tried. Redburn Atlantic double-downgraded McDonald's in June 2025 on an estimate that GLP-1 appetite suppression could cost the chain 28 million visits and $481.5 million a year, about 0.9% of system sales, and initiated Domino's at Sell on the same logic (43). FTI Consulting's ZIP-code analysis found quick-service wallet share in high-adoption, high-income ZIPs falling from 2.86% to 2.38% between 2022 and 2025 while full-service share rose from 7.68% to 9.05% (44). Those are the two most useful numbers for a site-level feasibility study, because they say the effect is geographic and it is a transfer between formats, not a uniform decline.

The SBA credit angle is where this becomes a lending question rather than an equity one. Fiscal 2025 was a record year for the 7(a) program at 77,600 loans and $37 billion, with another 6,750 504 loans for $7.8 billion (45). SBA publishes no official series for restaurant loans by NAICS code, which means anyone who wants one has to build it from the loan-level FOIA files, as MMCG has. In MMCG's own tabulation of the fiscal 2010 to 2019 cohort, 5.9% of full-service and 6.6% of limited-service restaurant loans were charged off, and franchise loans charged off at 9.4% against 7.1% for non-franchise (46). The FTC's 2024 analysis of 66,291 SBA franchise loans from 2013 to 2023 put restaurant defaults at about 5% and bars, taverns and nightclubs at 9%, with Dickey's Barbecue Pit at 20% (47). GAO found SBA paid guaranty claims on 28% of franchise 7(a) loans from fiscal 2003 to 2012 (48).

Restaurants, then, are already the highest-loss corner of the SBA book. A demand variable that shaves one to two points off quick-service revenue, concentrated in pizza, burgers and snacking dayparts, lands on loans with the thinnest cushion. No lender has written that down. We return to it below.

The packaged-food producers have the clearest view of the category data and the strongest incentive to blame something other than their own pricing. Their answer, with striking consistency, is that GLP-1 is real and small, and that what is actually hurting volume is three years of price increases meeting a consumer who has stopped absorbing them.

Frito-Lay North America, the largest salty-snack business in the country, saw volume fall 3% in the fourth quarter of 2024 and 2% in the third quarter of 2025 (49). PepsiCo's management attributed the weakness to "the cumulative impacts of inflationary pressures and higher borrowing costs on consumer budgets," difficult comparisons, and "continued growth in away-from-home dining and experiential spending." When CEO Ramon Laguarta addressed GLP-1 directly in February 2026, he said "there are more opportunities than threats, but they are both," and announced investment in single-serve capacity because 70% of the U.S. food portfolio already sells in single-serve formats (49). What PepsiCo did in the meantime is the part a lender should notice: it closed Frito-Lay manufacturing plants in Liberty, New York (287 jobs), Rancho Cucamonga, California (432) and Orlando, Florida (454) during 2025, with Laguarta's explanation that "the demand signal we had in '23 is different from the demand signal we have in '25" (50).

Hershey's filings tell the price story most plainly. North America confectionery volume fell approximately 5 points in the fourth quarter of 2025, 4 points in the first quarter of 2026 and 10 points in the second quarter of 2026, against net price realization of roughly 10, 12 and 14 points in the same quarters. Hershey attributed the volume loss to "price elasticity and normal quarter-to-quarter shipment variability" (51). Its former CEO called the GLP-1 effect "mild" in November 2024, and its current CEO has since pointed to gum and mints as a category that "benefits from functional snacking tailwinds, including GLP-1 adoption" (51). Mondelez's CEO Dirk Van de Put has been the most dismissive: "We're talking about 0.5% to 1% volume effect ten years down the road. I think the whole topic has been overblown," and in 2025, "there is currently no real impact on our volumes coming from GLP-1. It's all driven economically" (52). General Mills CEO Jeff Harmening put inflation first, said GLP-1 "has had a small impact so far," and then described the other side of the trade: Cheerios Protein approaching $100 million in sales within a year of launch, and a protein cereal portfolio generating roughly $200 million in retail sales and growing at a strong double-digit rate by September 2026 (53).

The product response has been fast and mostly cosmetic. Nestle's Vital Pursuit, the first national brand built explicitly for GLP-1 users, launched in September 2024 at $4.99 or under; Nestle later disclosed that 77% of its sales come from consumers who are not on the drugs (54). Conagra put a "GLP-1 Friendly" badge on 26 Healthy Choice items from January 2025 without changing a single recipe (55). Pop-Tarts and Doritos added protein, Pepsi launched a prebiotic cola. The measurable winners so far are protein dairy and protein cereal.

Which brings us to where the capital is going, because that is the part of this story that matters for an industrial feasibility study. While PepsiCo was closing salty-snack plants, Chobani committed $1.2 billion to a 1.4 million square foot dairy plant in Rome, New York, with up to 28 production lines and a stated intake of 12 million pounds of milk a day, broke ground in April 2025, and then committed another $1.2 billion to an Allentown, Pennsylvania plant for high-protein, low-sugar milk and protein shakes, alongside a $500 million expansion in Twin Falls, Idaho (56)(57). Coca-Cola's fairlife is building a $650 million, 750,000 square foot ultrafiltered milk plant in Webster, New York (58). The demand mix is moving from salt and sugar to protein, and the plant map is moving with it.

Alcohol is the category where the data are worst for the producers and the GLP-1 attribution is weakest. IWSR has U.S. beverage alcohol volume down 5% in 2025, with beer and wine each down 6% and spirits down 4%, falling in 49 of 50 states (59). The Beer Institute counted a 5.9% drop in domestic shipments, 8.68 million barrels, down in all 50 states (60). Gallup's drinking rate fell to 54% in 2025, the lowest in the survey's 90-year history, and stayed there in 2026 (61). IWSR's own explanation is affordability, and it added a line that undercuts the GLP-1 narrative: "the number of people drinking is not changing" (59). Diageo's CEO said the effect is "small when you think of spirits specifically," and the company's 2026 annual report records that early data show the impact is "lower on spirits when compared to beer and wine" (62). Constellation cut its beer guidance in September 2025 and blamed a "challenging macroeconomic environment" and Hispanic consumer weakness; its CEO has called the GLP-1 impact negligible (63). Molson Coors said in 2024 it had no data suggesting a meaningful effect (64).

For a feasibility analyst the alcohol story is a lesson in attribution. A category down 5% to 6% a year with a GLP-1 contribution the panels put at 1% to 2% of user-household spend is a category with a bigger problem than GLP-1. The drug is a tailwind on a trend that was already there.

The USDA angle completes the manufacturing picture. The Business and Industry guaranteed loan program obligated $1.60 billion in fiscal 2022, $1.86 billion in fiscal 2023 and $1.80 billion in fiscal 2024, with manufacturing at 18.5% of fiscal 2024 dollars and accommodation and food services at 32% (65). Through August 31, 2026 the program had obligated $1.01 billion of a $2.09 billion fiscal 2026 allotment (66). The Value-Added Producer Grant program awarded $26.5 million to 194 projects in September 2026, $11.6 million of it to beef (67). Rural food processing is a core B&I use, and a protein-dairy or meat-processing plant is exactly the kind of project a USDA lender sees. What the lender will not find is any USDA Economic Research Service or Food and Nutrition Service analysis of what GLP-1 adoption does to food demand. None exists. The only federal-adjacent work is a cooperative-agreement working paper that carries a disclaimer, and a Food Policy article arguing that ERS should buy the GLP-1 usage data and open it to researchers (68). The gap is itself a finding.

Every demand shift has a receiving end. For GLP-1 it is three places: the gym, the pharmacy counter and the protein aisle.

Fitness operators have said the loudest things and disclosed the fewest numbers. Life Time's CEO Bahram Akradi called GLP-1 drugs "a home run" for gyms and said betting that they would hurt the health club business is "a wrong bet," on a quarter in which average monthly dues rose 10.5% to $230 and revenue per membership rose 10.2% (69). Planet Fitness told investors a franchisee-commissioned survey found 50% of GLP-1 users consider a gym membership, launched a Ro partnership offering GLP-1 access through its Perks program in late 2025, and by August 2026 described it as "one of our most successful and most utilized Perks programs to date," on a base of 21.5 million members across 2,930 clubs (70). Equinox, Crunch and F45 have each built a GLP-1 program or partnership. The Health and Fitness Association counted a record 81 million U.S. facility members in 2025, up 5.2% (71), and William Blair's user survey found gym membership rising three points to 35% after starting the drug, with 72% of users reporting that they exercise more (25).

The clinical logic is sound, since rapid weight loss on a GLP-1 takes lean mass with it. The membership logic is less proven. Planet Fitness added just over 700,000 net new members in the first quarter of 2026 against about a million a year earlier, and its second-quarter same-club sales growth of 1.7% came entirely from rate (70). The strongest statement a feasibility study can make today is that GLP-1 adoption is a positive demand indicator for strength-oriented fitness formats, with evidence from operator commentary and user surveys and not yet from membership counts.

Pharmacy is where the money is measurable. Walmart disclosed that branded GLP-1 sales added about 100 basis points to Walmart U.S. comparable sales in each of fiscal 2025 and fiscal 2026, and guided the fiscal 2027 benefit to about half that as price cuts and Medicare maximum fair pricing bite (72). Costco's pharmacy sales grew nearly 20% in its fiscal fourth quarter of 2026 on its GLP-1 program and lower GLP-1 pricing (73). CVS reported pharmacy same-store sales up 18.0% in 2025, "primarily driven by pharmacy drug mix, including branded GLP-1 drugs" (74). The telehealth prescribers have ridden the same wave with worse economics: Hims & Hers posted $753 million of second-quarter 2026 revenue on 2.9 million subscribers, with gross margin down to 64% from 76% as it pivoted from compounded to branded supply (75), and WeightWatchers, which went through Chapter 11 in 2025, now gets a quarter of its revenue from 197,000 clinical subscribers (76). For a pharmacy-anchored retail center or a weight-management tenant, that is a demand story with real numbers behind it and a regulatory risk attached: the FDA issued 30 warning letters to telehealth compounders in March 2026 and 25 more in June.

Convenience retail is where the operators disagree with each other, which makes it the most instructive case. Alimentation Couche-Tard's CEO attributed weakness in salty snacks, carbonated soft drinks and confectionery "partly" to GLP-1 drugs in September 2026 and said shelf space would be reallocated to protein and functional products (77). Casey's CEO Darren Rebelez, reporting the same quarter, said snack weakness was "primarily pricing-driven," pointed out that national-brand chips fell 8% in units while Casey's private-label chips grew 16%, and added: "If it was a GLP-1 impact, I don't think we'd see the strength in our own private brand" (78). NACS data support both of them. Convenience-channel salty snack sales fell 11.5% in the first half of 2026, while alternative snacks (jerky, seeds, nuts) rose 7.9% in 2025, a shift NACS explicitly attributed to protein demand from GLP-1 users (79). The honest reading is that the center-store snack decline is mostly price and private-label substitution, with a GLP-1 overlay that is changing what sells more than how much.

The sixth research pass asked a narrow question: has any lender, rating agency, appraiser, CRE research house or federal forecaster incorporated GLP-1 adoption into credit analysis, valuation or demand projections for restaurants, food manufacturing, convenience retail or fitness? As of October 5, 2026, the answer is no.

The quantified work lives entirely in equity research and strategy consulting. Barclays is the one bank that framed it as a credit trade, recommending in October 2023 that clients buy protection on PepsiCo, McDonald's and Altria because "credit market prices don't fully reflect the potential downside" (80). Piper Sandler's 45 to 110 basis point annual packaged-food drag is a sales estimate, not a credit metric (30). Moody's held a webinar on the topic in June 2024 and has published nothing since. S&P Global's midyear 2026 industry credit outlook mentions GLP-1 only as a growth driver for pharmaceuticals, and its restaurant commentary attributes pressure to costs and inflation (81). Fitch has nothing. No CBRE, JLL, Cushman & Wakefield, Green Street, Trepp or CoStar research ties GLP-1 to restaurant tenants, net-lease QSR values or food-industrial demand. National asking cap rates for net-lease quick-service restaurants in the first quarter of 2026 were 5.82% for corporate and 6.80% for franchisee credit, with McDonald's ground leases at 4.40%, and no source attributes a basis point of that to GLP-1 (82). FRANdata's unit success rate for franchising is projected to decline to 94.2% in 2025 from 95.9% in 2021 on macro drivers, with no GLP-1 variable (83). SBA has published no guidance, notice or portfolio data that mentions it. Neither the Appraisal Institute nor any other professional body has issued methodology. USDA ERS, the Bureau of Labor Statistics, the Congressional Budget Office and the Federal Reserve have no demand-side treatment of it anywhere in their published work.

The effect is new, it is entangled with a larger price-driven traffic decline, and the best evidence is eighteen months old. But a lender who finances a pizza franchise in a high-income suburb in 2026 on a pro forma that cites 2019 traffic benchmarks is carrying a risk that the data can now size, and is not sizing it. That is what a feasibility study is for.

A feasibility study does not need to predict the adoption curve. It needs to answer three questions about one project in one trade area: how exposed is this market, how exposed is this format, and what does the exposure do to debt service coverage if it runs at the measured rate. MMCG now builds that as a named sensitivity in every restaurant, franchise, food-manufacturing, convenience and fitness study, layered on the base case rather than replacing it. The method has three inputs.

The first is market exposure: the share of households in the trade area that are GLP-1 households. The national figure is 21% to 23%. The local figure moves with income, sex, age and region. Adoption runs highest among women and among adults aged 40 to 64 (1)(3). In the weight-loss segment it climbs with income, and claims data put patients earning over $250,000 at 72% more likely to be on a GLP-1 than those under $50,000; geographically it concentrates in the South, Appalachia and parts of the Midwest where obesity and diabetes prevalence are highest (84). In MMCG's demographic model, a trade area with a $120,000 median household income and a 45-year-old median age is a 28% to 32% market, and a rural Mountain West trade area is 12% to 15%. The FTI ZIP-code work shows the quick-service wallet share shift is already visible in the high-income, high-adoption ZIPs and not elsewhere (44).

The second is format exposure: the per-household effect for the kind of business being underwritten, taken from the matched-control panels. For limited-service food that is 8% to 9% of spend at six months, attenuating toward 4% to 5% at twelve months as discontinuation and reversion work through. For pizza it is double that. For full-service dining the panel effect on visits is neutral to slightly positive and the effect sits in alcohol and attachments. For salty-snack and confectionery manufacturing it is 7% to 11% of category spend. For protein dairy, nutrition bars and strength-format fitness the sign is positive.

The third is persistence, which converts a household effect into a sector effect. About a third of users discontinue within the study window and revert (19)(7), so the steady-state effect for a household on the drug is roughly 60% to 70% of the six-month peak, and the sector drag is the market exposure multiplied by that figure. For a quick-service restaurant in a national-average trade area, that is 22% of households times roughly 5% to 6%, or a 1.1% to 1.3% revenue drag against the no-GLP-1 counterfactual. In a high-adoption trade area with a pizza format, it is 30% times 15%, or 4.5%. Those figures reconcile with the bank estimates: Piper Sandler's 45 to 110 basis points a year for packaged food, Morgan Stanley's 1.6% of calories by 2035, OC&C's 0.2% a year on U.S. food and beverage volume through 2031 (85).

What that does to coverage is the only number the lender needs. Take a limited-service restaurant with $3.0 million of stabilized revenue, $450,000 of cash flow available for debt service and $360,000 of annual debt service on an SBA 7(a) loan, a 1.25x coverage ratio. Restaurant flow-through on lost revenue runs about 50% once food cost falls away and labor is partly flexed. A 2% GLP-1 revenue drag is $60,000 of revenue and $30,000 of cash flow, and coverage falls to 1.17x. A 4% drag, the high-exposure pizza case, is $60,000 of cash flow and coverage of 1.08x, below the 1.15x most SBA lenders require at stabilization and within reach of the 1.00x line in a bad quarter. The project does not fail on GLP-1 alone. It fails on GLP-1 stacked on a 3% traffic decline from gas prices and a rent step, which is the point of running a sensitivity: it shows how much of the cushion a single structural variable consumes before the cyclical ones arrive.

The same arithmetic runs in reverse for the beneficiaries. A strength-format gym in a high-adoption trade area gets an upward sensitivity on membership demand, flagged as operator-evidenced rather than panel-evidenced. A protein-dairy or value-added meat processing plant financed under USDA B&I gets a demand case that the category data support and that no federal forecast contradicts, because none exists.

The table below is the exposure matrix MMCG now carries into asset-class studies. Direction is the sign of the GLP-1 sensitivity on the base case. Channel is where the effect shows up in the operating statement. Evidence grade is panel (matched-control transaction data), survey (self-reported), operator (company disclosure) or inference (category data without direct GLP-1 attribution).

Three caveats travel with the table. The ranges are steady-state against a counterfactual, not year-over-year forecasts; a market that was growing 3% still grows 1% to 2% under the quick-service case. The high-adoption trade area multiplies each range by roughly 1.4. And the whole table assumes the measured effects of 2024 and 2025 hold as the user base shifts toward Medicare patients and pill users, who may consume differently; the 2027 panel data will settle that.

A feasibility study submitted in 2026 for any asset class in the table above should carry four things that a 2024 study did not. A trade-area GLP-1 exposure estimate, built from the demographic ring rather than assumed at the national average. A format-specific revenue sensitivity drawn from the matched-control panels, with the source and the window stated. A coverage test that stacks the GLP-1 sensitivity on the study's existing downside case rather than running it alone. And, for a franchise, a note on whether the franchisor has named the drugs in its 10-K or FDD risk factors, since two major quick-service franchisors already have.

None of that changes the determination on most projects. On the thin ones it does, and the thin ones are the ones that end up in the charge-off data.

Does GLP-1 adoption actually reduce restaurant traffic?

The best transaction data say it reduces spending at limited-service restaurants by about 8% per adopting household in the first six months, with the cut falling more on items per order than on visit counts. Full-service and casual dining spending among the same households is flat to slightly up after a year. Chains report no measurable traffic effect because it is buried inside a larger price-driven decline.

How many Americans are on a GLP-1 drug in 2026?

Surveys put current use at 11% to 12% of adults, about 29 to 32 million people, and roughly one household in five contains a user. Prescription claims show about 11 million, because they miss compounded and cash-pay supply. Both figures are rising with the arrival of pills, price cuts and Medicare coverage.

Which food categories lose the most?

Chips and savory snacks, sweet bakery, frozen convenience sides, cheese, cookies, soft drinks and pizza, with household spending cuts of 7% to 22% depending on the category and the study. Yogurt, fresh produce, nutrition bars and meat snacks rise modestly.

Is the alcohol industry decline caused by GLP-1?

Mostly not. U.S. alcohol volume fell 5% in 2025 and the drinking rate is at a 90-year low, but the household panels put the GLP-1 effect at 1% to 2% of user-household alcohol spend, and IWSR reports the number of drinkers is unchanged. GLP-1 is a tailwind on a decline driven by affordability and health trends.

Have any lenders or rating agencies priced GLP-1 into restaurant credit?

No. As of October 2026, no rating agency, CRE research house, SBA guidance, appraisal methodology or federal forecast incorporates GLP-1 as a demand variable. The quantified estimates exist only in bank equity research and consulting work.

How does MMCG include GLP-1 in a feasibility study?

As a named sensitivity on the base case: trade-area household exposure, multiplied by the format-specific effect from matched-control panel data, adjusted for persistence, and run through debt service coverage stacked on the study's existing downside case. For beneficiary asset classes the sensitivity runs upward and is graded by evidence quality.