BBB Foods runs roughly 3,600 ultra-discount grocery stores in Mexico, and management now says the country can hold more than 14,000 of them.
The usual way to size a hard discounter is against the supermarket chains it is attacking. Leandro, who writes the Best Anchor Stocks newsletter, says that misses where the share is actually coming from: about half of Mexico's grocery market is not supermarkets at all, but self-employed people running small informal shops.
"I thought that I was not being extremely conservative but this business when it works it can produce numbers that seem a bit unrealistic."
Leandro publishes in-depth company and industry research at Best Anchor Stocks, has owned BBB Foods for less than a year, and has already revised his own model upward twice because the company outran it.
I listened to the full interview so you can skip it. 62 minutes of audio, 21 minutes of reading.
Here are the 12 takeaways that matter.
👤 Guest: Leandro, who writes Best Anchor Stocks, a Substack newsletter of in-depth research on high-quality and niche companies
🎙️ Host: Brett Schaefer, co-host of Chit Chat Stocks, a CCM Media Group podcast
📰 Published: 9 September 2026 on YouTube (Chit Chat Stocks Podcast)
🔴 YouTube | 🟣 Apple Podcasts | ⏱️ 62 min | ✅ Time saved: 41 min
Key Takeaways
About half of Mexico's grocery market is informal, and that is the share BBB Foods is taking
The same setup existed in Turkey, where the hard discounter BIM industrialized it first
The founder copied a Turkish business he watched from the inside at Merrill Lynch
He looked for a country with similar conditions and picked Mexico
Same-store sales grew 18% last quarter with 3,600 stores already open
Management's Mexican store opportunity has moved from 12,000 to above 14,000
The reason to own a hard discounter is cash conversion, not margins
Suppliers get paid after the inventory has already turned several times
Leandro says the company has no direct competitor in Mexico right now
Incumbents have to cut their own margins to respond, and public-company chief executives rarely will
Reported operating margin went from 2% to negative 1%, and it is a stock-compensation artifact
He uses adjusted EBITDA and models the dilution separately, to avoid counting the same cost twice
The stock is up 50% and his expected annual return is still above 15%
Earnings grew faster than the share price, so the multiple barely moved
Robotics needs 100 times the computing power of a large language model, on Jensen Huang's math
The likely robotics winners are companies that already won in digital AI
Pure-play robotics names had already risen 300% to 500% before he started looking
Analog chips are the piece that joins the digital and physical worlds
Texas Instruments' chief executive put the content at roughly $1,000 per humanoid
1. Copying BIM From Turkey
Leandro found BBB Foods through a friend who already owned it and told him he would like the business. The friend also handed him the framework: a book called "Retail Disruptors," about how the hard discount model ate into the market share of established grocery chains.
The book's best-known case is Tesco losing share to the German discounters. Leandro relayed it with a hedge: "I think they invested in Tesco and Tesco started to lose significant share against the hard discounters from Germany, Aldi and Lidl"
What the book gave him was a durability argument, not a growth one. "So it basically explains how the hard discount model works and why it has let's say durable advantages, right? Which in retail is not easy to find"
Anthony Hatum, the founder, president and chief executive, previously worked in Merrill Lynch's private equity arm, which held an investment in BIM, a Turkish grocery hard discounter. Impressed by BIM's numbers, he went looking for somewhere to copy it: "Well, maybe I should go to find a place where I can replicate this model"
He settled on Mexico and opened the first store in the early 2000s — Leandro said he thought 2004 but explicitly did not want to be held to the year
The company now does around $4 billion in annual revenue, on Leandro's figure, and is still growing fast
Leandro's framing is that the country choice was not incidental. Mexico was picked because its grocery industry looked like Turkey's, which is why he says the company should not be compared with a hard discounter competing in a developed market
2. How Hard Discount Works
Asked what a store actually looks like, Leandro described small formats in dense urban neighborhoods, designed for customers who walk.
The model runs on deliberately few products. "Because the idea of a hard discount model is that you simplify as much as possible so that you drive a lot of volume per SKU. So you can get a lot of cost savings per SKU and you share those cost savings with the customers"
As volume builds, those products get replaced with the company's own private-label versions, which are cheaper for the shopper and carry a higher margin for the retailer
Shopping frequency is high and baskets are small: a typical customer goes two or three times a week, and Leandro put the average ticket below $5, hedged as a figure he did not know by heart
He splits Mexican grocery into three tiers, and says the middle one loses. Luxury grocers carry many brands and services to justify higher prices; low-cost operators run a shared-scale model that passes savings on; and, in his words, "everything in the middle has ended up disrupted"
He also named a higher-end Mexican supermarket chain that he says is doing well despite the rise of hard discounters, as evidence that the top of the market is defensible too
3. Half The Market Is Informal
This is the part Leandro says investors in developed markets miss. Mexican grocery divides not only into price tiers but into a formal, industrialized market and an informal one — self-employed people running small shops in densely populated areas.
The informal share is roughly half the market, and it is largely unable to fight back. He described the informal operators as not even organized among themselves, which he said means there is very little they can do in response
Turkey looked the same when BIM started: Leandro put the informal share of the Turkish grocery market at 60% at the time, made up of small informal shops
"What BIM did in Turkey was basically industrialize the informal market," and he says the same thing is happening in Mexico now
On why Hatum picked the country: "I mean the two countries are very similar in the sense that the informal market made around half of the total grocery market, right?" There was no hard discounter disrupting either the formal or the informal side
The competitive consequence is that the company does not need to beat the incumbents to succeed. He thinks it is gaining share from them as well, but says that is not what the thesis rests on
4. Not A Dollar Store
Brett Schaefer asked what makes the model fit Mexico rather than the United States, and both of them spent time on how differently the two countries shop.
Leandro tied store economics to walking distance: a small store in a dense area can serve a lot of people, and of high-frequency, small-ticket shopping he said: "So, basically that's not a setting where you would go by car if you can avoid it because you are going to be able to take the food walking to your house." Management has added parking where it judged it necessary, but he does not consider parking part of the value proposition
He put Europe in between, with both small neighborhood supermarkets and large drive-to stores, and characterized the US as hypermarket territory: "I mean, I go to Costco in Spain. So, I can feel that's more the US model than when what we typically get here"
Schaefer pushed to head off a US reader's assumption: "This isn't like a Dollar General or a Dollar Tree. It's a lot different because I think when people think, oh, small concept store, ultra cheap prices, they're like, well, is this just a dollar store in Mexico?" Leandro's answer: "It's all groceries"
The distinction he draws is that a dollar store sells convenience and this sells price. A dollar store is closer to your house than the cheaper option; the hard discounter is both closer and cheaper. And on the dollar stores themselves, he noted that "pretty much everything costs more than a dollar"
The company does not yet sell produce or meat, a point Schaefer raised. Leandro said fresh is now being piloted, and explained the constraint: "So when you include a new category, they typically take another one that's not performing well out, right?" Pilots run in a limited number of stores, and if the returns are high the category goes network-wide — but something else comes out
5. 3,600 Stores, 14,000 To Go
Schaefer put the end-of-Q2 store count at around 3,600 and asked what the runway looks like.
Management's own Mexican store opportunity has been raised. In Leandro's account management "began at 12,000 in store opportunity" and has since taken it above 14,000, a figure he flagged he was giving from memory
The unusual part is that same-store sales are growing at the same time. "I think in the last quarter same store sales grew 18% or something like that," and he said roughly half the quarter's total growth came from existing stores rather than new ones
Schaefer's own numbers from the show: same-store sales have averaged around 16% over the past five years, and store count has compounded at around 30% a year over the same period. He said he could not think of another business producing 16% comparable-store sales
Leandro's explanation for the organic growth is brand density — as more stores open in an area, more people know the chain and widen what they buy there
He checks management's target against Turkey rather than taking it on faith. "You can go to BIM in Turkey to see what their store density ended up at and when you do that exercise you end up in a relatively similar number"
On expanding outside Mexico, he is dismissive for now. BIM has already gone into Morocco and Egypt, but it is a far more mature business. Of international expansion for BBB Foods: "I wouldn't say it's even in the top 10 priorities of the company right now." His view is that it belongs in a durability case rather than a growth case — "But at this point in time, I wouldn't take the international opportunity seriously to be honest"
6. Cash Beats Margins Here
Schaefer noted gross margins hovering around 16% and asked where they settle at scale. Leandro's answer was that margins are the wrong thing to focus on.
The company could take the scale economies for itself and does not. "At the end it's like I'm not going to discover here anything new this is scale economy shared the Amazon model the Costco model all over again"
His model assumes gross margin rises from the 16.5% he expects this year: he said "I have higher gross margins that 16.5%" in his own model
On mature EBITDA margin he anchors to BIM, at mid-to-high single digits, and said his model uses roughly 7% to 8% — while flagging that reported margins are trending better than he assumed
"The best thing about this business is not the margins, right?" The point is cash conversion. Scale wins better supplier terms, and private-label suppliers end up selling almost exclusively to the chain, because it is not a bad position to be in, "serving a retailer that's growing a 40% clip"
Few products, fast rotation: "So they turn the inventory several times before paying suppliers," which he contrasted with grocery retailers that turn inventory more slowly
The BIM precedent is that the model funds its own expansion and still throws off cash. Over the decade from 2015, Leandro said, BIM funded all of its store expansion internally and generated free cash flow worth about three times its starting market capitalization — "So they generated three times their initial market cap a decade ago while not needing to raise one single dollar to continue growing"
7. No Direct Competitor Yet
Asked to name the largest competitor, Leandro gave an answer he acknowledged sounded odd.
"I don't think they really have a direct competitor right now," he said, splitting the competitive set into incumbents and the informal market
The incumbent problem is structural: responding to a hard discounter means cutting your own margins, and a listed company's chief executive is usually unwilling to do that. "That's what happened to companies like Tesco until eventually it's too late and then you react. But when you react these hard discounters are already large enough that you cannot destroy them"
He relayed a former employee's view from an expert call: Walmart could have killed the company cheaply in its first years, and now, at $4 billion of revenue, cannot. That is his account of someone else's claim, not his own analysis
Once you are the low-cost operator, he said, the first mover can "keep the gap with competitors constant" because its share and its scale keep rising
Other small-format chains exist but are built differently: "There are other concepts in the country that are also trying to be hard discounters but a lot of them are more convenient stores than hard discount stores." Schaefer named Oxxo. Leandro added that Walmart runs a competing format too — "Walmart also has a concept that's trying to compete with, I think it's called Bodega Aurrera"
The durability argument rests on grocery not being fashion. "Because trends are unlikely to change. People are going to still need to eat in 10, 20 years time"
The risk he names is self-inflicted. The main danger, he said, is that the company stops turning the flywheel: "That you start doing things that are not what a hard discounter would do and then you give an opportunity for competitors to go in"
Schaefer contrasted this with Aldi's US push, where the competitive set is far denser: "I mean, Aldi definitely has more of a competitive an intense competitive set in the United States where they're trying to expand to"
8. Reading Through The SBC
Schaefer flagged a chart that will confuse people. On reported operating margin: "It's gone from 2% over the last 12 months back in 2024/2025. Now it's down to negative 1%."
Leandro's answer is that the reported figures are distorted by post-IPO stock compensation. "One has to be careful with the reported figures," he said, noting that management does disclose how that compensation flows into the fully diluted share count
His method avoids double-counting the same cost: "I use the adjusted EBITDA figures and ignore the SBC expense and then I run it through dilution"
On capital intensity, he builds it from store openings. Current pace is around 400 to 500 store openings a year, which he thinks can scale to about 1,000 a year, because the regional structure copied from BIM decentralizes real-estate decisions. "BIM in Turkey topped around 1,000 stores per year"
Distribution centers scale with stores: "I think the average has been that they currently have around 170 stores per distribution center." Store count and that ratio give the number of distribution centers, and construction costs for both give the capital expenditure line
The income statement makes it look capital intensive and the cash flow statement does not. "On a cash flow basis the business is not really that capital intensive because you're getting the benefits of the significantly negative working capital." Suppliers are paid after the goods have sold, so growth is funded by the model: "You're basically getting free financing from the model"
9. Grocery Before New Bets
Schaefer asked whether non-grocery services — financial services of the kind Oxxo runs in Mexico — belong in the model. Leandro's answer echoed his position on international expansion.
There is still room inside plain grocery. Fresh is a large category the company does not have yet, so "there's still ample runway within grocery" before it needs anything else
Anything new has to fit the discount model. Management is willing to "stay very close to the hard discount model," so an added category needs high turnover, genuine customer demand, and ideally the ability to be private-labeled
On the most recent earnings call the chief executive described the business as a platform rather than a grocery retailer. Leandro's gloss: "I mean, if you have a lot of traffic from customers, that gives you a lot of optionality to offer things customers want, right?"
The company already rotates non-grocery categories through the stores to create a treasure-hunt effect — items appear for days or weeks and then go. "Maybe you can find TVs there and then they'll disappear and they'll rotate that category"
He ranks the two optionality cases against each other: "I'd say even that it's significantly higher probability that we'd see non-grocery over the next couple of years than that we would see an international expansion" — but says neither is required
10. Up 50%, IRRs Above 15%
On valuation, Leandro warned that the screens are misleading and that recent price action has not made the stock expensive.
The headline multiples are broken by the stock compensation. "If you look at the reported figures, it's going to look hyper expensive especially on profitability metrics due to the SBC"
He also cautions against current or forward-year multiples on a business growing this fast, and noted the stock was "trading around one time sales not long ago"
The share price and the business have moved together. "The stock has run up like 40% in no time like in six months but at the same time the business has grown 40%," so the multiple has barely moved
His valuation method has three inputs: revenue from store openings plus same-store sales, a mature EBITDA margin, and the fully diluted share count. The margin assumption dominates, because "changing it 100 bips up or down" — 100 basis points — moves the valuation a lot — and he uses BIM as the proxy for where the financials end up, a comparison management itself has made
On the share count, he said stock compensation is used to align employees who are also shareholders, and that a further dilution allowance has to be added on top of the reported fully diluted number before computing EBITDA per share
The return math is the striking part. "When I added the business to my portfolio, I was getting say 17-18% IRRs in my model. The stock is up 50% and my irRs are still above 15%," because earnings grew faster than he expected. He added: "I've have to update my model to the upside a couple of times already and I've not even held the company for one year"
The lock-up expiry on 6 August did not do what he expected. Rather than a rush of selling, the stock went to all-time highs. "Everyone thought that after the lock up expiration, everyone would rush to sell their shares, but Anthony Hatum said that he didn't expect people would rush to sell their shares when they are growing 40% while expanding margins"
His conclusion on price: "I wouldn't say it's very expensive. Evidently, it would have been better to buy 50% lower like it always is, but the business has continued to grow fast"
He used it as a case study in price anchoring: "You buy and then you're up 100% in say two years or three years and you say whoa I like paid half for this so now it's expensive but then you check at the fundamentals and maybe the business has grown even above that 100% and it's even cheaper today." Schaefer's summary: "It's a good problem to have, but it can be uncomfortable when you have a stock do too well"
11. Robotics Is Further Off
The last third of the interview covered Leandro's sector report on robotics. He came to it through a three-stage view of AI.
The stages, in his framing: large language models and generative AI first, then agentic AI, which he places in its early innings, and then robotics, which he described as "taking agentic AI to the physical world"
He wrote the report to learn: "And writing is a great way to force yourself to understand especially when you're writing and sharing it publicly," because readers who know the subject will correct you
He deliberately wrote it in plain terms. "I find that in this kinds of topics, a lot of people write very complex articles just to feel that they like to make others feel that they know a lot"
His main conclusion is a timing one: "The main takeaway for me was that we are still a bit far off from a full on robotics wave"
The reason is two constraints instead of one. Digital AI is limited by computing power, which is in turn limited by electricity. Robotics adds a second. It is far more compute-intensive than digital AI, he said, and on top of that you have to "physically manufacture a lot of things." Software scales fast; factories do not. "So I think it's still a bit far away or maybe more far away than people expect"
The market has already moved anyway. Niche suppliers and pure-play robotics names had already risen 300% to 500% in a very short period before he started looking. "So I was surprised to see that a lot of that is already being anticipated in the stocks," he said — "I know that the market is forwardlooking but one could also argue that the market can be too much forwardlooking in some cases right and maybe anticipated too soon"
He splits robotics into three markets: industrial robotics, the most mature, where automation is long established and Amazon runs robots in its warehouses; autonomous driving, which he thinks the industry is "sort of starting to solve" now; and humanoids, the least developed. Humanoids follow the usual technology adoption curve: "Especially because like in every technology when you are early on probably it's very expensive to manufacture at scale and when you want adoption you need to bring the cost down"
12. The Winners Already Won
Leandro's second robotics conclusion is that the obvious way to play the theme is not the best one.
What robotics actually is, financially, is a demand shock for computing. "Because basically what underpins robotics is a huge wave in computing demand"
The scale, on Jensen Huang's math as Leandro relayed it: "I think Jensen Huang used to say that if you think about LLMs as 1x in computing, then agentic AI is 10 times that requirements and then you have 100 times in robotics, right?" Robots reason in the physical world, where the cost of failure is high
Demand only pays shareholders where the supplier is differentiated. "If you get a huge influx of demand but there's no differentiation then the entire value is going to accrue to the customer or to the end consumer." He thinks the chip ecosystem does have that differentiation in many places
Nvidia is running the same playbook it ran in digital AI. "The strategy that Nvidia is following with robotics is pretty similar to what they've followed in digital AI, right?" It supplies the software and model tools, then monetizes the computing that follows: "So if you use Nvidia tools to build robotics then those tools are going to work better with Nvidia hardware right so you are going to end up buying Nvidia hardware"
Schaefer raised Texas Instruments, which is weighted toward industrial and automotive customers. Leandro said analog chips are what join the two worlds: "So, I think the Texas Instruments CEO sized the opportunity, and my numbers may be off here, by around $1,000 per humanoid when they come," with the same executive cautioning that humanoids will not scale up suddenly in the next few years
Analog is already benefiting from the digital AI build-out through power management, and he thinks the segment has competitive advantages — but "I must say that a lot of companies don't look precisely cheap today after the run"
On what investors get wrong, he named the timeline first. People are "maybe a tad too optimistic on the timeline" in robotics, when digital AI is already hitting physical limits on build-out speed and electricity
The second error is looking too hard for the niche name. Hunting for the small supplier that goes parabolic is high risk and high reward, because it is too early to know who wins — while nobody asks whether the companies already working will benefit. "But Nvidia is going to probably be a strong winner in robotics and nobody is thinking about it." Same for Texas Instruments: "Do a lot of people think about Texas Instruments when they think about robotics? Probably not"
He conceded why: robotics has to get much bigger to move the numbers at Nvidia or Texas Instruments, perhaps 10 to 20 years out, whereas with smaller pure-play robotics names "you basically are levered towards the topic"
Bonus Insights
The book that framed the whole thesis was hard to buy. "Interestingly, it's not on Amazon. Well, at least not when I purchased it. So, you have to buy it on the web page"
Leandro's definition of a quality business has shifted over his career: "I must say that my investment philosophy has been molded through the years. So my definition of quality today is not the same as it was 3 four years ago"
Schaefer said Leandro is probably the show's most frequent returning guest over the past five to six years, and that listener numbers are why they keep booking him
Leandro said before recording that he had been researching companies neither host had heard of, which Schaefer used to make the point about how little-covered his coverage universe is
BIM's underlying growth in Turkey, stripping out currency depreciation, has run "upwards of 20-30% for many many many years," on Leandro's account
Asked what it feels like to hold a stock that has run this hard, he said: "I would love to have this problem with all of my stocks"
Leandro's bottom line is that BBB Foods is early in a hard discount rollout whose real competition is Mexico's informal shops rather than Walmart, which is why he still models expected annual returns above 15% after a 50% run — and that the cleanest way to own the robotics build-out is through the semiconductor companies already winning in digital AI, not the pure-play robotics names that have already tripled.
Products, Companies & Tools Mentioned
BBB Foods (TBBB) (The Mexican hard discount grocer at the center of the interview — roughly 3,600 stores, around $4 billion of revenue, and management's stated opportunity for more than 14,000 stores in Mexico)
BIM (The Turkish hard discounter the founder studied at Merrill Lynch and copied; Leandro uses it as the template for store density, mature margins and cash generation)
Walmart de México (The listed incumbent Leandro says could have killed the company early and now cannot; it runs the competing Bodega Aurrera format)
Oxxo (The Mexican convenience chain Schaefer named as the nearest small-format comparison, and the model for non-grocery services such as financial services)
Tesco, Aldi and Lidl (The case study from "Retail Disruptors" — Leandro's account is that Tesco lost significant share to the two German hard discounters)
Nvidia (The robotics name he thinks is mispriced by omission: it supplies the software and model tools, then monetizes the hardware that follows)
Texas Instruments (Analog chips as the bridge between digital and physical AI; its chief executive sized the content at roughly $1,000 per humanoid)
Costco (Leandro's example of the American drive-to hypermarket model, which he says does not describe how Mexicans shop)
Dollar General and Dollar Tree (The comparison Schaefer raised and Leandro rejected — dollar stores sell convenience, a hard discounter sells price)
Amazon (Cited twice: as an operator already running robots at scale in its warehouses, and as the shared-scale model the grocer is copying)
Books & Resources Mentioned
Retail Disruptors (The book a friend recommended to Leandro, on how hard discounters took share from established grocery chains; he said it was not sold on Amazon when he bought it)
The Durable Winners of the Robot Age (Leandro's robotics sector report, the basis for the second half of the interview; part of it is free to read)
Best Anchor Stocks (Leandro's newsletter — in-depth company reports, earnings follow-ups and industry write-ups; he has also published on semiconductors and the alcohol industry)
If this was worth your time, send it to someone who follows the name.
Get the latest market chatter as it happens: