Chipotle Q1: comps positive, AI talk modest, and the 32% loyalty number that should worry every fast casual

I read the Chipotle and Brinker prints back-to-back on Wednesday morning, both filed before the open, both on the same desk, both pulled up on the same screen. It was a quiet morning — no Fed, no jobs print, no mega-cap tape running over the top — so I had the rare luxury of actually sitting with the numbers for an hour before the sell-side notes started landing. The Chipotle release hit my inbox at 4:05pm ET the prior afternoon in the wire pre-read and then officially this morning; the Brinker 8-K earnings release for Q3 FY26 hit at 6:45am ET. Two fast-casual-adjacent prints on the same Wednesday, both with the same uncomfortable message for the AI-premium crowd that has been bidding up restaurant tech multiples since the back half of 2025.

The contrarian thesis I want to put down before the analyst day decks lock the narrative: AI is not yet the lever for restaurant operators — operations are, and loyalty depth is the real moat. Chipotle ran +0.5% on the comp with positive traffic, +7.4% on revenue to $3.1B, and a loyalty program now sitting at 32% of sales — up 300 basis points year-over-year. Brinker ran +4.0% at Chili’s, the 20th consecutive quarter of positive comp growth, and lifted adjusted EPS to $2.90 from $2.66 prior. Neither one of those two stories is being driven by a marquee AI announcement this quarter. Both of them are being driven by execution. That is the story that the desk notes are going to bury under the EPS-miss headline at Chipotle and the consensus-beat headline at Brinker, and that is the story investors actually need to internalize before they price the next fast-casual tech roll-up.

The Chipotle numbers, plain and unadorned

Let me set the Chipotle Q1 down in one place so the math is in front of us.

Revenue printed $3.1B, up 7.4% year-over-year. Same-store sales were +0.5%, with transactions up 60 basis points and average check up roughly 10 basis points — the rare clean fast-casual quarter where the comp is being driven by traffic rather than price. Restaurant-level margin came in at 23.7% on the adjusted basis (after backing out the legal settlement charge that hit the GAAP line). Adjusted diluted EPS was $0.24, down 17% year-over-year. Openings were 49 net new units, 42 of which were Chipotlanes — the drive-thru pickup format that has been the unit-economics workhorse for three years running. Loyalty enrollment crossed the threshold that matters: 32% of sales now flow through the rewards program, up from 29% a year ago.

The EPS miss is the headline the wire is running. The miss is real — $0.24 versus the $0.28-ish consensus depending on which compiler you trust — and the proximate causes are beef inflation (the carne asada and barbacoa input cost ran high single digits on the quarter) and freight (still elevated from the Q4 weather disruption that lingered into January). Scott Boatwright, the CEO, framed the cost line on the Q1 transcript as transitory and the loyalty mix as structural, and I think the framing is correct — but the framing is also what every CEO says when the cost line is bad. The thing the transcript actually rewards on a careful read is not the cost-line defense, it is the loyalty narrative. Boatwright spent more minutes on rewards engagement and personalization than he did on any of the AI projects — Ava Cado included.

The 32% number and why it should worry every fast casual

Here is where the math gets uncomfortable for the rest of the category.

Chipotle’s loyalty program is now sitting at 32% of sales. The fast-casual peer set — and I am drawing the line broadly here to include Cava, Sweetgreen, Wingstop, Chicken Salad Chick, and the regional bowl chains that have been pitching deals at me for the last six months — is sitting somewhere between 5% and 15% of sales on rewards mix. Cava is the closest at roughly 22% in their last disclosure. Sweetgreen is in the low-to-mid teens. Wingstop is high single digits because their model has been call-and-pickup for years and they only recently formalized a rewards layer. The regional bowls I am seeing in dealflow are reporting 8%-12% loyalty penetration on average, and the smaller they are, the lower the number.

What does the 32% number do for Chipotle that the 8%-12% number does not do for everyone else?

It does three things, and all three are durable.

One: it compresses the cost of incremental traffic to near zero. When a third of your sales come from identified, app-engaged customers, the marketing spend per incremental visit collapses. The CFO line item that used to be national-television advertising becomes targeted push notifications and tier-based offers. A fast casual running at 8% loyalty mix is still paying for traffic the old way — paid social, paid search, occasional broadcast. A fast casual running at 32% is paying for traffic with stored data and a CRM. That gap is worth somewhere between 80 and 150 basis points of restaurant-level margin over a full cycle, and it compounds.

Two: it gives you a real-time read on demand elasticity that you cannot buy. When you raise price by 2%, the 32% loyalty mix tells you within 72 hours whether your top quintile of customers blinked. The 8% mix tells you within six weeks, by which point the elasticity decision has already been made. The Chipotle pricing discipline of the last four years — measured, well-paced, never panicked — is a function of having the data to know when to push and when to hold. Most of the peer set is still flying blind.

Three: it makes the AI investments meaningfully cheaper on a unit-of-output basis. This is the point that I think the Chipotle AI stack post lays out in detail — when you run AI personalization against a 32% loyalty mix, every model dollar gets amortized across roughly four times the addressable transaction volume of a peer running 8%. The same Ava Cado AI Operator that posted Boatwright’s appointment in March is running on top of that loyalty base, and that base is the reason the AI investments will eventually move the P&L. The investments themselves are not the lever. The base is.

Brinker: the same Wednesday, the opposite framing

Now flip to the Brinker print, which landed three hours before the Chipotle release on the same Wednesday morning.

Brinker reported Q3 FY26 revenue of $1.470B, adjusted diluted EPS of $2.90 versus $2.66 in the prior-year quarter, and — the line I underlined twice — +4.0% comp at Chili’s, marking the 20th consecutive quarter of positive same-store sales growth. Twenty quarters. Five full fiscal years. Through a pandemic, through 2022 inflation, through the 2024 fast-casual share-shift narrative, through every consumer-spending wobble in between. The Chili’s turnaround that Kevin Hochman has been running since 2022 is no longer a turnaround story — it is a structural execution story.

What is the Chili’s AI announcement that has driven the 20-quarter run? There isn’t one. There is no Chili’s Ava Cado. There is no Chili’s AI Operator. There is no proprietary LLM in the kitchen, no agentic loyalty platform, no machine-vision quality control at the line. The Chili’s story is the “Big Smasher” burger launch, the Triple Dipper menu architecture, the back-of-house labor model that runs at lower hours-per-cover than the casual-dining peer set, and the marketing discipline of a team that figured out their customer is in the $50,000-$80,000 household-income band and stopped trying to be everything to everyone.

That is the story the AI-premium thesis has the hardest time digesting. A casual-dining brand with no AI narrative just ran 20 consecutive quarters of comp growth and printed a 9% EPS beat on a Wednesday morning, while a fast-casual brand with one of the most-discussed AI stacks in the category just missed EPS by 14% on the same morning. The difference is not the tech. The difference is the operating muscle and — in Chipotle’s case — the loyalty base that is doing the structural work while the AI investments mature.

The bet: operational discipline plus loyalty depth, not the announcement cycle

Where does this leave the M&A book?

It leaves me, frankly, on the same page I have been writing for the last six months across the case against the AI premium, the M&A roundup, and the Four Margins framework. The AI announcements are interesting. Some of them will eventually move the P&L. None of them are the lever today, and pricing them as if they are is how mid-market fast-casual deals are getting overbid by 1.5-2.0 turns of EBITDA.

The bet I am willing to underwrite right now, in this market, on this set of comps, is the same bet Chipotle and Brinker are both already running. Buy operational discipline. Buy loyalty depth — and by loyalty depth I mean the Paytronix-Starbucks deal architecture that pulls rewards mix from single-digits to high-teens inside 18 months on a real loyalty CRM. Buy unit economics that work at the current cost structure, not at the cost structure that the AI roadmap promises in 2028. Buy management teams that talk about traffic and check separately on the earnings call rather than smearing them together into a comp number.

The 32% loyalty mix is the number to anchor on. If a target is running at 8%-12%, the diligence question is not “what is the AI roadmap.” The diligence question is “what is the realistic path to 25% in 36 months, and what is the marketing spend curve that gets us there.” That is the bet. The AI story rides on top of that base or it does not ride at all.

The cost line will come back. Beef will normalize. Freight will normalize. The 17% adjusted EPS decline at Chipotle is a cyclical cost-line story sitting on top of a structurally improving traffic story and a structurally deepening loyalty mix. That is a buyable setup at the right multiple, and the right multiple is not the AI multiple — it is the operations multiple. The two are diverging, and the divergence is the trade.

— Oliver writes The Bottom Line for TableTransfers. Tips: ma@tabletransfers.com.