Darden +4.2% comps and the boring Bahama Breeze ending

I was on the print at six-forty-five Thursday morning with Luca standing over the laptop and a flat white going cold next to the trackpad. The Darden 8-K dropped, the deck loaded, and the first thing I did — the thing I do every Darden cycle — was Ctrl-F for the letters A and I. I got three hits in the prepared remarks, all incidental. The number I was looking for was not there. The numbers that were there are the story.

The contrarian read — and this is INTERPRETATION, not anything Darden said — is that the best old-line full-service print of the March cycle is the one with the least AI in it. Rick Cardenas spent the call talking about retention, portion architecture, and a LongHorn culinary recertification program for managers. He did not pitch a forecasting tool. He did not name a kitchen-AI vendor. He delivered +4.2% blended comps and let the absence do its own work.

The print, plainly

Per the Q3 FY26 results exhibit filed with the SEC Thursday morning: total sales of $3.3 billion, up 5.9% year-over-year. Blended same-restaurant sales of +4.2%. LongHorn Steakhouse comps of +7.2%. Olive Garden comps of +3.2%. Adjusted diluted earnings per share of $2.95. The company repurchased $127 million of stock in the quarter. The fiscal year benefits from a 53rd week, which the release names directly; the comp number is reported on a comparable basis, and the comp number is the one that matters.

The PRNewswire release carries the dividend declaration and the updated fiscal 2026 outlook. The CEO quote in the release is the one to mark: Cardenas saying “all segments grew sales and segment profit dollars.” Segment profit dollars. Not segment margin percentage, not a basis-point story — dollars. Read against the Q3 FY26 earnings call transcript, where Cardenas and CFO Raj Vennam walked through brand-by-brand, the sentence is doing a specific job: telling the buy side that nothing in the portfolio is leaking. Even Bahama Breeze, which has been the segment of the portfolio analysts spend the most time worrying about, ends the quarter unremarkably. Boring is the word. Boring is the goal.

A note on what is reportable and what is interpretation. The sales, comp, EPS, and buyback numbers are in the release. The Cardenas commentary is on the call. Everything in the next two sections is interpretation, and I will mark it as such.

Why TableTransfers cares

This is the INTERPRETATION section. Darden just produced the cleanest counter-example to the AI cost-takeout pitch this publication has covered all winter. The pitch — the one Oliver will dissect in his forthcoming buy-side note on the AI premium, the one David will frame against in his Four Margins essay — is that legacy full-service operators recover margin by deploying kitchen-AI, scheduling-AI, and voice agents on top of an unchanged concept. Most of the pitch, as I have written before, is real on its own terms. The case studies are real. The contracts are real.

Darden did not need any of it to print +4.2%. What Cardenas named on the call instead, in order of how much time he spent on it: record team-member retention, a “lighter portion” section that Olive Garden rolled into the menu and is reading well with guests, and a culinary recertification program for LongHorn managers — every operating partner re-trained on the steak, the seasoning, the plate. That is not a technology story. That is an execution story, run by a CEO who has been in the seat long enough to know which levers in a casual-dining P&L respond to attention and which do not.

The contrast with the Cracker Barrel Q2 print on March 5 is the one to sit with. Cracker Barrel went into its call with a -10.1% traffic line and a tariff problem and no concept-side answer. Darden went into its call with retention at a record, a menu architecture that gives guests a value option without discounting the rest of the check, and a steakhouse brand whose managers can all cook the protein to the same standard. Same segment, same macro, completely different print. The variable that explains the gap is not technology. It is operating discipline that has been compounding for three years.

What I will be watching

Three things, through the Q4 print and into FY27.

One: whether LongHorn’s +7.2% holds. A culinary recertification program is a one-time event; the lift it produces is durable only if the operating partners stay re-trained and the protein keeps hitting the plate the same way. If LongHorn prints +5% or better in Q4 without a recertification narrative attached, the program graduates from initiative to culture, which is the most expensive thing a competitor can try to copy.

Two: whether anyone on the Darden bench introduces a named AI deployment in the Q4 or Q1 FY27 call. My read is that they will not, and that the absence is deliberate. Cardenas is not anti-technology. He is anti-narrative. The moment Darden names a vendor, the story stops being “we executed” and starts being “we are betting on the same thing everyone else is betting on.” The premium evaporates.

Three: the $127 million of buybacks. At an adjusted EPS of $2.95 in a quarter where blended comps printed +4.2%, the capital return is the tell. Management is saying, with the balance sheet, that the best use of incremental cash this quarter was their own stock — not a tech acquisition, not a vendor contract, not a deployment program. Mark that. Operators do not buy back stock at this pace if they think the next twelve months require a transformation.

The Q3 print is good. The Bahama Breeze ending is boring. The AI footnote is the part to reread.

— Luca covers restaurants for TableTransfers. Tips: tips@tabletransfers.com.