DoorDash Q4: Deliveroo and SevenRooms Are Integrated. 2026 EBITDA Guidance Is the Bill.

The release hit at 4:05 p.m. Eastern. I was at the desk in the print, the photographer was still packing lights two feet to my right, and the numbers landed on the Bloomberg side panel before I had finished my second coffee. Orders 903 million, +32%. Marketplace GOV $29.7 billion, +39%. Revenue $4.0 billion, +38%. Adjusted EBITDA $780 million, +38%. Free cash flow $254 million (DoorDash IR).

First read on the desk: clean print. Every line green. The Deliveroo close in October and the SevenRooms close in June had compounded the marketplace into 40+ countries on a single P&L for the first time, and the consolidated quarter put the integration-cost story to bed for anyone who was worried the deals would break the comp.

The stock opened the after-hours session down ten percent.

The print was not the story

Sell-side clustered around $3.99 billion in revenue and $1.29 EPS. DoorDash reported $4.0 billion (a $31 million beat) and $1.27 (a two-cent miss). Orders and GOV both beat. By the rule that the print drives the first hour and the guide drives the next twelve, the print was a positive surprise.

Marcus filed his pre-earnings setup last Friday in The Bottom Line. The read going in was that the consolidated top line would print clean, that the integration disclosure was the thing to watch, and that the cost-rolloff schedule would be the question for the May Q1 call. He had the print roughly right. He had the reaction roughly wrong — he expected the integration disclosure itself to be the swing variable. What actually moved the stock was three sentences inside the 2026 guidance section.

The guide is the bill

DoorDash guided Q1 2026 Marketplace GOV to $31.0–$31.8 billion and Q1 adjusted EBITDA to $675–$775 million. The Q1 EBITDA midpoint is the soft spot — at $725 million it sits below the Q4 print of $780 million, and below where consensus had Q1 marked.

Then the full-year line. For 2026, DoorDash said adjusted EBITDA as a percentage of Marketplace GOV is expected to be “up slightly” versus 2025 — excluding the Deliveroo impact (DoorDash IR). That parenthetical is the bill. Strip Deliveroo out of both years and the underlying business expands margin slightly. Put Deliveroo back in and the consolidated 2026 margin is, by the maths of the disclosure, flat to slightly down versus 2025. SBC $1.3–$1.4 billion. D&A $1.1–$1.2 billion. The cost of running two parallel tech stacks is sitting in there.

CFO Ravi Inukonda walked the analysts through it. The cleanest version, per the trade write-ups that ran after: 2026 EBITDA margin will be up slightly compared to 2025 excluding Deliveroo, and redundant costs from parallel platforms are concentrated mostly in 2026 before rolling off (tikr). “Mostly in 2026” is the phrase to underline. Not “entirely.” A smaller portion bleeds into 2027.

What Tony Xu said about Deliveroo

The Deliveroo question was whether the acquired business is tracking the May 2025 underwriting case. The close priced in a specific cost-synergy path, and the question for the next eighteen months is whether Deliveroo’s underlying growth is fast enough to absorb the integration spend without the consolidated margin going negative on the year.

Xu’s read, per the trade transcripts: Deliveroo is “gaining share in its largest markets.” That is the answer the deal model needs. The +39% consolidated GOV bakes in Deliveroo for a full quarter for the first time, and the underlying Deliveroo growth rate sitting inside that number is what’s giving Xu the runway to spend through 2026 without renegotiating the integration timetable with the board.

The sub-text: SevenRooms is smaller in dollar terms and already further along. Going Out launched in September. The DashPass-tied reservation layer is live in Miami and New York and expanding through 2026. That work hits OpEx but doesn’t require the multi-year parallel-stack carry that the Deliveroo migration needs. The 2026 EBITDA bill is mostly a Deliveroo bill.

Why the stock moved ten percent

Three reasons, descending order.

One. The guide reset the 2026 EBITDA bar. The market had been carrying a 2026 number that assumed integration costs were front-loaded into 2025 — the close-year — and would roll off cleanly in 2026. The disclosure said the opposite: the parallel-stack carry is concentrated in 2026. Real reset, not a noise move.

Two. “Up slightly excluding ROO” is the kind of guide language that asks investors to separate the underlying business from the deal year. Some did. Some did not. The ones who did not read the headline guide as a margin print and sold.

Three. The two-cent EPS miss gave the after-hours wires the on-paper number to write the negative story against. It would not have moved the stock on its own. Combined with the 2026 framing, it was the line that ran.

What this means for operators

If you’re on SevenRooms. The integration cost is the company’s bill, not yours. Going Out is live in your launch market or is coming to it in 2026. Your renewal should anchor on the data-rights questions the SevenRooms M&A read walks through — what crosses the line between SevenRooms and the DoorDash commerce platform layer, and what your guest-data export rights look like if you ever walk.

If you’re on Deliveroo’s old international stack. You are about to be migrated onto the DoorDash unified platform on a schedule that has not yet been published. The migration is what the 2026 EBITDA bill is paying for. Ask your account rep what the cutover looks like on your account, and what the rollback path looks like if menu, pricing, or dispatch breaks during the cut.

If you’re reading the DoorDash stack as a competitive map. The Q4 print confirms the read in The Pass on DoorDash’s commerce platform — this is a hospitality technology company that happens to also run the largest delivery marketplace. The 2026 EBITDA guide is the company telling you, in dollar terms, how much the platform consolidation is costing this year.

The watch for May

The Q1 2026 call is May 6. Two lines to watch: the consolidated EBITDA margin print versus the Q1 guide midpoint, and whether Inukonda re-words or re-affirms the “mostly in 2026” framing. If the wording tightens — if “mostly” becomes “concentrated in” or pairs with a 2027 rolloff range — the 2026 reset will be the trough. If it loosens, the bill is bigger than tonight’s print suggested.

Tonight’s read: the deals are integrated, the platform is consolidating on schedule, the underlying business is growing into the multiple. The bill is sitting in the 2026 guide, and the market is going to argue about it for the next two quarters. The print was clean. The guidance line was the one to read.

— Maya covers restaurant tech for The Pass. Tips: maya@tabletransfers.com.