Cava's 1,000-Store Math: AUVs Above $3M and the Unit-Economic Case for a Path No One's Walked

I was on a call with a Cava franchisee — sorry, a Cava regional operator; the brand is corporate-owned and that distinction matters in a minute — when the Q2 release hit the wire this morning. He had the press release open in one tab, the SEC 8-K in another, and a spreadsheet that he had been maintaining since the IPO in a third. The spreadsheet had a column for what he called “the 1,000 number.” Every quarter, he updates two cells: trailing-twelve-month new-store AUV, and the comp from the most recent print. From those two cells, he computes a single output: years until the unit count has to actually be 1,000 if the company means what it says.

This morning, after he read the Q2 release, the number got better, not worse. That sounds wrong if you only read the headline. The headline today is that Cava’s same-store sales rose 2.1% in the quarter, the slowest comp the company has printed as a public entity, and the stock — depending on which intraday tick you’re looking at — is down something like 20% as I write this. The narrative being written about Cava today is a narrative of deceleration. The narrative I’m going to write is different. It is that 2.1% on top of the comps Cava has been stacking is exceptional inside a category that is, for most of its members, negative right now, and that the operative number for anyone trying to underwrite this brand is not the comp at all. It is the AUV on the 2025 class. That number, per the company’s own disclosure, is now above $3M. If that number holds — and it has held longer than skeptics have given it credit for — the math for 1,000 units works. If it doesn’t, nothing else in the model matters.

This piece is for operators who are trying to decide what to believe about Cava in particular and what to believe about the broader fast-casual-Mediterranean thesis in general. I’m going to walk through what Q2 actually said, what the 2025-class AUV implies for the build-out economics, why the comp deceleration is not the story, why the cost line items moved the way they did, what could break the thesis, and what an operator at another concept should steal from how Cava is communicating right now.

What Q2 actually said, in plain language

The earnings release is unusually clean for a public restaurant filing. The big numbers, in the order an operator reads them:

Revenue of $278.2M, up 20.3% year-over-year. Same-restaurant sales up 2.1% — and inside that 2.1%, traffic was modestly negative, with the balance carried by check, which is mostly mix and a small amount of price. Net new restaurant openings of 16 in the quarter, bringing the total to 398, which is one short of the round number the company will cross during Q3 and which the press has spent a remarkable amount of energy underweighting because 399 doesn’t headline. Restaurant-level profit margin of 26.3%. Net income of $18.4M, which is in fact, as AInvest’s coverage flagged, a year-high in dollar terms despite the EPS print being lower year-over-year because of share count and tax dynamics. And the guidance: 68 to 70 net new restaurants for the full year, same-restaurant sales of 4.0% to 6.0% for the full year — which means the back half is implicitly expected to step up from the 2.1% in Q2 — and the path-to-1,000-by-2032 language, which was reaffirmed.

Two numbers that aren’t in the release headlines but are in the body and in the prepared remarks: the 2025 class of new restaurants is running at an AUV above $3M, and trailing-twelve-month AUV for the system is approximately $2.9M.

Mark interpretation: the $3M-plus class AUV is the single most important number in this release. It is not even close. Everything I am going to say about the path-to-1,000 thesis is downstream of whether you believe that number is durable or whether you believe it is going to fade. The comp, the guidance, the margin — all of those are second-order. The first-order question is: does a new Cava restaurant, opened in 2025, in the markets Cava is opening it in, produce more than $3M of sales in its first year, and does it keep producing $3M-plus in its second, third, fifth, eighth year?

The 1,000-store math, done out loud

Here is the unit-economic question reduced to a back-of-envelope an operator can do in their head.

If you are going from 398 restaurants to 1,000 restaurants over roughly seven years, you need to open net 600. At the guided cadence — 68 to 70 net new this year, which is roughly 17% unit growth — you can get there if the cadence holds and the absolute number of openings ramps as the base grows, because at 17% unit growth on a larger base you produce more openings per year in absolute terms even if the percentage tapers somewhat. The cadence isn’t the problem. The problem is the unit economics underneath the cadence.

For a restaurant company to commit publicly to opening that many units, the units have to clear an internal rate-of-return hurdle that justifies the capital. The build cost on a new Cava restaurant is, per company disclosure, somewhere in the high $1M range fully loaded — call it $1.5M to $1.8M for math. If your AUV is $3M and your restaurant-level margin is in the mid-20s — and Q2 just printed 26.3% at the system level — then a new unit produces something like $750K to $800K of restaurant-level profit per year. Against a $1.5M to $1.8M build, that is a payback in the two-year zone and a cash-on-cash return that puts the unit, mathematically, in the top decile of public-restaurant unit economics. There are a handful of concepts that historically have produced returns at this altitude. Chipotle in its 2014–2018 build years. Chick-fil-A perpetually, although Chick-fil-A doesn’t publish numbers and operates a different model. Texas Roadhouse in its sweet spot. Wingstop on the franchisee side, although that’s a different cost structure entirely. Cava, if the $3M-plus class AUV holds, joins that list.

If, however, the $3M class fades to the system $2.9M — and then fades further as new units open into less-cherry-picked markets — then the math gets squishier. At $2.6M of AUV with 24% restaurant margins, you’re producing $624K per unit, and against a $1.7M build you’re looking at a payback closer to three years and a return that is still good but no longer top-decile. At $2.4M with 22% margins, the model still works but it doesn’t justify the multiple the stock is being asked to support. So when I say the $3M number is the only number that matters, I mean it in the most literal way. The discounted future cash flow of this entire enterprise turns on whether new-class AUVs stay above $3M as the geographic footprint expands.

Why the 2.1% comp is the wrong number to anchor on

I want to be careful here because if you go too hard the other way — “ignore the comp, only watch AUVs” — you sound like every IR deck that has ever been published. So let me be specific about why I think the comp deceleration, in this print, in this category, in this quarter, is not the operative signal.

First, the base. Cava’s comparable-restaurant-sales stack going into Q2 2025 was historically extraordinary. The company has been compounding sales per restaurant at a rate that essentially no public restaurant company has matched in this cycle. Comping +2.1% on top of that stack is, in absolute dollar terms, still a meaningful per-restaurant gain. The hurdle was simply enormous.

Second, the category. The fast-casual segment broadly is running negative-low-single-digits in comp right now. Chipotle’s most recent print was negative; the QSR composite is negative; the casual-dining composite is more negative than that. In a category that is, by any honest read, contracting in real terms, a +2.1% comp at a brand that is also opening 17% new units annually is not deceleration. It is outperformance.

Third, the composition. Inside the +2.1%, traffic was modestly negative and check carried the quarter. The reflex for an analyst reading that line is to assume price-driven check growth, which is bad — it implies you’re squeezing existing customers. But the prepared remarks attributed most of the check movement to mix, particularly to the rollout of grilled steak as a permanent menu item, which carries a higher attach rate and price point than chicken or falafel. That’s a different quality of check growth. Mix-driven check growth from a successful protein expansion is, structurally, more durable than price-driven check growth, because customers are opting into the higher-ticket choice rather than being pushed into a higher ticket by menu reprices.

Fourth, and this is the part the market is most underweighting today: guidance for the full year was reaffirmed at 4.0% to 6.0% same-restaurant sales. That implies the back half steps up materially from the 2.1% printed in Q2. Either management is wildly wrong about its own back half — possible, but not the base case eleven days into Q3 — or there is a specific reason they believe Q3 and Q4 reaccelerate. Reading between the lines of the prepared remarks, that reason is the lapping of last year’s grilled-steak launch and the rollout of the next LTO cycle into the holiday window.

So: 2.1% on top of a historic stack, against a category in negative territory, with mix-driven check quality, and a credible reacceleration path baked into reaffirmed guidance. That is not the deceleration story the stock is telling today. That is something else.

The margin line, and why it didn’t break

The 26.3% restaurant-level margin held up better than most analysts modeled going into the print. CNBC’s coverage flagged the margin resilience as the surprise of the print, against a backdrop where everyone in the fast-casual space has been telegraphing labor and food inflation pressure all summer.

The breakdown matters. Food, beverage, and packaging as a percent of sales ticked up modestly, driven by commodity inflation on a basket that for Cava is heavy on proteins, dairy (tzatziki, feta), olive oil, and fresh produce. Labor as a percent of sales held roughly flat year-over-year, which is remarkable in a year when the federal minimum-wage debate has gone nowhere and most operators are seeing 4–6% wage inflation on the line. The lever that has historically held labor in check at Cava is throughput per labor hour — the make-line is designed to move more guests per hour per labor minute than the comparable Chipotle line, and as AUVs climb, that ratio improves further. Occupancy as a percent of sales declined, which is the natural effect of comp growth on a largely-fixed rent line. Other operating expenses moved within range.

Mark interpretation: 26.3% is not the high-water mark for this brand. The Q1 print was higher. The Q2 step-down is partly seasonal and partly a function of commodity timing. The number to watch through the back half is whether the brand can hold 25% with the back-half comp acceleration the company is implicitly guiding to. If they can, the model produces full-year restaurant-level margins comfortably above the 24.8% to 25.2% range most analysts had been carrying. If they can’t, the bear case picks up a tooth.

What about the broader Mediterranean thesis?

There is a separate question hovering above the Cava-specific debate, which is whether Mediterranean fast-casual is a real category that supports multiple billion-dollar brands or whether it is a Cava-shaped market in which Cava is essentially the entire category. I think this is the most underrated question in fast-casual right now, and I do not think Q2 settled it.

The Cava-only camp’s argument is straightforward. The brand has no meaningful direct competitor at scale. Sweetgreen is not a direct competitor; it is a salad concept that has historically operated at a higher price point and is now leaning hard into its Infinite Kitchen automation thesis — a forthcoming May piece walks through how the unit economics on that platform shift the picture. Chipotle competes for the same lunch occasion but operates a fundamentally different cuisine. The smaller Mediterranean players — Roti, Naf Naf, Verts — have not scaled at the rate that would suggest the category supports multiple brands at Cava’s altitude. By this read, Cava is sui generis, and the 1,000-store path is justified because there is no one to take the units from them.

The category-is-real camp’s argument is also defensible. Mediterranean cuisine indexes well on every consumer-trend variable that matters for the next decade — perceived health, perceived freshness, protein density, ingredient familiarity, customization-friendliness — and the absence of a strong second player at scale is, in this read, a temporal accident rather than a structural fact. Under this read, the next decade produces at least one well-capitalized challenger and probably two, and the question for Cava is whether they build the moat — through real estate, through brand, through digital — fast enough to remain dominant when the challenger arrives.

I lean toward the second read, but it doesn’t materially change the 1,000-store math, because even in a world with a credible challenger, the addressable market for Mediterranean fast-casual probably supports 1,000 Cava units and 500 challenger units without saturation. The U.S. has roughly 8,000 Chipotle units, more than 13,000 Starbucks units, more than 2,000 Chick-fil-A units. A 1,000-unit Cava is, in unit-count terms, a mid-size fast-casual presence. The geographic gating constraint historically has been real estate quality and brand awareness in unproven markets, not category saturation.

The four-margins frame, applied

For readers who haven’t seen it, the four-margins framework — an upcoming May framework piece — argues that every restaurant brand operates against four distinct margins that have to be managed in concert: restaurant-level margin (what the unit produces), corporate-level margin (what the enterprise produces after G&A), real-estate margin (the spread between site cost and site productivity), and brand-margin (the premium consumers will pay versus alternatives). Most brand narratives collapse to one or two of these. The honest read of any concept requires looking at all four.

Cava’s four-margin picture, post-Q2:

Restaurant-level margin: 26.3%, top-decile for the category, with structural advantages from throughput-per-labor-hour and from menu-mix on grilled steak.

Corporate-level margin: less clean. G&A is still being absorbed by a unit base that is, by the standards of the path-to-1,000 endpoint, small. As units scale, G&A as a percentage of sales should compress. The model becomes meaningfully more profitable at 600 units than at 400, and meaningfully more profitable still at 800.

Real-estate margin: the part of the story that does not get enough air time. Cava’s real-estate playbook has been disciplined — they have not chased trophy locations the way Sweetgreen did in its early years, and the brand-led rent negotiation leverage is real, because Cava is now a desirable co-tenant in lifestyle centers and high-density mixed-use developments. The $3M AUV class is partly a real-estate-quality story.

Brand-margin: the part of the story that gets overstated. Cava’s price point is in the $13-$15 ticket range for a bowl with the steak upcharge, which is more expensive than Chipotle but less expensive than the salad concepts. The brand premium is real but not enormous, and it cannot be the load-bearing wall of the thesis. The thesis has to be supported by the unit economics under the brand premium, not by the brand premium itself.

Under this four-margin lens, the picture is: very strong restaurant-level, leveraging-into-strong corporate-level, structurally-strong real-estate, modestly-positive brand. The combination produces a credible 1,000-store thesis. The combination would not produce that thesis at lower restaurant-level margins or lower real-estate productivity.

What breaks the thesis

I want to be honest about the bear case, because if I just write the bull case it isn’t journalism.

The thesis breaks if the 2025 class AUV fades. This is the bear case in one sentence. If you open the 2026 and 2027 classes into less-favored markets — and the geographic footprint expansion path implies exactly that — and those classes produce $2.5M instead of $3M, then the marginal unit economics no longer justify the build cost, and the path-to-1,000 either stalls or has to be financed at returns below the cost of capital. The leading indicator for this would be a quarter where management discloses class-level AUVs and the most recent class is below $2.9M trailing-twelve-month. We are not at that quarter yet. We may never be. But it is the watch item.

The thesis also breaks if a credible challenger arrives faster than the moat builds. I think this is less likely than the AUV-fade risk, but it is non-zero. The category is attractive enough that a well-capitalized entrant — picture a private-equity-backed roll-up of regional Mediterranean concepts — could plausibly assemble a 200-unit competitor inside three years. Whether that competitor would have Cava’s brand, ingredient sourcing, throughput discipline, and real-estate position is another question, but the option exists.

The thesis also breaks, in a smaller way, if the labor-line discipline I described above breaks. Cava’s labor efficiency is a function of make-line design and software-supported throughput optimization. Both of those are imitable. If competitors close the throughput gap, Cava loses one of the structural advantages that has produced its margin profile.

None of these break-points are imminent. The watch items are forward indicators, and Q2 did not move any of them in a worrying direction. But they are the things I would put on the dashboard if I were underwriting this brand as an operator.

What to steal from how Cava is communicating

A note for operators reading this who are not on the Cava cap table: the communication discipline in this print is worth studying separately from the financial substance.

The release led with what the company wanted you to know, not with what would have been easiest to write. Revenue and unit growth are the framing; the comp deceleration is acknowledged in the body without being apologized for; the guidance is reaffirmed clearly; and the class-AUV disclosure is treated as load-bearing rather than buried in a footnote. The prepared remarks did the same thing: management addressed the comp question directly, explained the back-half acceleration thesis without overselling it, and put the $3M class AUV in front of the conversation rather than waiting for the analyst to dig for it.

The lesson for operators communicating about their own concepts — whether you’re public or whether you’re talking to a private board or to a lender — is that you control which number anchors the conversation by choosing where to put the emphasis. Cava chose the AUV. The market is choosing the comp. Whether the AUV anchor ends up winning is, in part, a function of whether the AUV holds, but it is also a function of whether the brand keeps putting the AUV in front of the conversation. They will. They should. Most brands wouldn’t.

What I’m watching from here

Through the back half: whether the comp accelerates into the 4.0% to 6.0% guided range, and whether the 2025 class AUV trajectory shows up in any Q3 commentary in a way that confirms or weakens the $3M number. Through 2026: whether the new-class AUV stays above $3M as the openings cadence moves into geographies that are less proven for the brand. Through the back half of the decade: whether competitive entrants emerge and how Cava’s brand moat performs against them.

The 1,000-store number is, today, a credible target. It was not a credible target two years ago. It is credible today specifically because the unit-economic line items have moved in the direction the path requires. The brand has earned the right to make the commitment.

The market disagrees this morning. The market is going to be wrong before it is right and right before it is wrong; that is what markets do with growth-stage restaurant equities. The operator question, separated from the equity question, is whether the unit economics described in this release are the unit economics that justify building the next 600 stores. My answer, after walking through the numbers on the call this morning with the regional operator and his spreadsheet, is yes. The number that matters is above $3M and holding. Everything else is noise inside that signal.

— Priya covers operators for TableTransfers. Tips: [email protected].