Caught in the middle.
Every operator is now two companies.

In the second quarter of 2026, Walmart de México opened 21 stores. Seventeen were Bodega Aurrera Express, its smallest and cheapest box. It opened zero Supercenters, zero Walmart Express and zero Sam's Clubs — every new unit fell inside its discount banner. Chedraui opened 28 stores of which 27 were Supercito, and closed two, both Súper Chedraui. Across Walmex's whole network, 23 openings added roughly twelve thousand square meters of selling floor, an average of 522 square meters per store, against an installed Supercenter averaging over seven thousand.

Read the column of zeros. Sam's Club was simultaneously Walmex's best-performing format of the quarter, with double-digit growth in member acquisition and more members upgrading to the Plus tier — and it received no new units. A company does not stop building the format that is working unless the constraint is something other than demand: the land, the anchor tenancy, the catchment area, the weekly car trip that fewer households are still making. Sam's wins on productivity, not on expansion. Bodega Aurrera Express wins on both, and takes all seventeen units.

The closures are more legible still. Súper Chedraui, Soriana Súper, Soriana Mercado. Three banners, two companies, one format — the mid-sized conventional supermarket, between two and five thousand square meters, wide assortment, priced above the discounter and below the club. Nobody in this universe closed a proximity store. Nobody closed a club. The mortality is concentrated in exactly one box.

FEMSA runs the two poles as separate companies, and this quarter it made that structural: it reorganized its reportable segments to give OXXO México a line of its own. OXXO added 253 net stores against 309 a year earlier, and 832 over twelve months against 1,218 — down 31.7%. Bara added 112 in the same three months to reach 786, a store base up 47.5% year over year, described by the chief executive as a record quarter for new store openings. Bara's 786 stores are a rounding error next to OXXO's 24,708. Its growth rate is not, and neither is the direction of the two curves.

And OXXO's own quarter shows what it costs to hold the line. Same-store sales grew 9.5% and traffic turned positive at 2.0% after several quarters of decline — a genuinely good result. But gross margin contracted 70 basis points, from 45.5% to 44.8%, and the company says why: price rationalization in traffic-driving categories, and affordability strategies across most categories to stay competitive against other channels. That last phrase is OXXO naming hard discount without naming it. Three things bought that traffic: price, a Spin Premia tender that reached 50.4% of sales, and the tournament — of which one has already expired.

La Comer shows the same movement over a decade rather than a quarter. The flagship banner has gone from 71% of the portfolio in 2016 to 50% today, surrendering twenty-one points to a premium urban format at one end and a proximity format at the other. The average La Comer store is roughly 6,800 square meters; the average Sumesa is under 800. The only investment the company highlighted this quarter was the Ps.80 million relaunch of Sumesa on June 3, positioned explicitly for urban proximity. The premium operator is running the identical play as the discounter. It is just further along.

The 2.4-point gap is new floor. The sector is adding square meters while its existing stores sell less — the industry-wide version of what La Comer and Chedraui show in sales per square meter. La Comer grew sales per square meter 0.9% and Chedraui 0.7%, against food inflation of 4.85%. The two operators that added the most selling floor built roughly four points of real negative productivity into it.

Price without a proposition — the cleanest test of the quarter

Two Coca-Cola bottlers. Same product, same excise tax, same quarter, same tournament. One raised price and lost units. The other conceded real price and gained them.

Everything above can still be read as a story about money: households under pressure trading down to the cheapest box available. The café sector is where that reading breaks, and it is the reason this issue argues the change is in the supply of options rather than in the size of the wallet. Nobody stops going to Starbucks because they are hungry.

A cohort of formats is expanding that has nothing in common by category and everything in common by structure. Tim Hortons opened its 200th Mexican store this year at Parque Fundidora in Monterrey; it had 65 units in September 2023 and 150 as recently as September 2025 — roughly fifty in the last twelve months. Sede Café runs 48 stores and 470 employees, built in two years from a single location that opened in February 2024. It is not a franchise, sources its coffee from the Istmo de Tehuantepec, and in the company's own description rarely has tables, because the format is built for the trip between point A and point B. Its most recent opening was in Vallejo, not Roma or Condesa. Pick Up Coffee runs about 50 takeaway-only units in Mexico City. El Globo, owned by Grupo Bimbo, is adding café service across its network.

Sede describes itself through three pillars: quality, accessibility, and bonito. Bueno, bonito y barato — the same three words that give Tiendas 3B its name. Two companies in unrelated categories, no common ownership, both growing double digits in the same market on the same three-word formula. Neither is selling the same product cheaper. Both delete the attributes the customer stopped paying for — the table, the name written on the cup, the wide assortment, the destination trip — and price what is left. It is the identical move Supercito, Bodega Aurrera Express, Bara and Sumesa make in retail, executed in a category where nobody was looking for it. ANTAD confirms the direction with the only format it has in positive territory: specialized retail, narrow propositions, +3.1% in June and +5.4% in July, while self-service and department stores fell.

The incumbent under attack is inside this universe. Starbucks in Mexico is operated by Alsea, whose cafeterías segment fell 2.0% in the quarter — the weakest of its three Mexican formats against full service at +5.4% and QSR at +2.1%, and the only one in negative territory. Deflated, that is close to seven points of real contraction. Alsea reports the segment rather than the brand, so the figure covers its coffee portfolio and not Starbucks alone; the direction is not in question. Starbucks is not losing the market — it remains the largest coffee chain in the country by a wide margin. What it is losing is the exclusivity of the occasion.

Its answer is the Coffee House Uplift, a program to improve the physical store. Consider what that response assumes. Nobody remodels the room when the problem is the price of admission; you remodel when you believe the customer would come if the room were nicer. But the attackers have removed the room entirely. Sede rarely has tables. Pick Up Coffee is takeaway only. A meaningful share of Tim Hortons' Mexican network is drive-thru. The incumbent is investing in the single attribute its competitors decided to stop charging for, on the theory that the attribute is why people came. Its own house contains the alternative: in the same portfolio and the same quarter, Vips grew on value platforms. One Alsea brand renegotiates the price of entry; the other refurbishes the place the customer is leaving.