DoorDash made money again.
The company reported on its Q4: 903M orders (+32% YoY), Marketplace GOV $29.7B (+39%), Revenue $4B (+38%), GAAP net income $213M (+51%).
1/ Agentic commerce question. An analyst asked Tony Xu – co-founder and CEO – whether AI agents could compress the discovery-to-checkout flow and erode DD’s position. Tony gave a bit of a history lesson: Amazon took product search from Google by solving the whole job: browsing, reviews, tracking, delivery, returns, support. Google launched Google Food Ordering in 2016, drove more traffic to restaurants than DoorDash could, and still failed – traffic didn’t convert or retain, because after checkout things happen in the physical world (oh, yes). A driver is late, an item is missing, a substitution needs to be made. The end-to-end job wasn’t solved. So Tony basically says AI agents become the new top-of-funnel, driving discovery same as Facebook and Google. The transaction and execution still land somewhere that owns the physical data, operations, personalization.
2/ DashPass. Ravi Inukonda, the CFO: record year, record quarter in Q4. Margin per order goes down, but profit dollars go up – subscribers retain more, order more. I posted a while back about DoorDash’s engineering work on promotions – a system that figures out which customers would order anyway and stops giving them discounts, getting the same incremental orders at half (!) the cost. DashPass seems to apply the same logic at the subscription layer: instead of handing out delivery fee waivers to people who’d order regardless, you charge a flat monthly fee upfront. Stop subsidising behaviour that was already going to happen.
3/ Expanding beyond restaurants. 30% of DoorDash MAUs already order from grocery, convenience, or retail. Ravi’s goal is 100% – every user eventually ordering across all verticals, not just restaurants. Uber reported 40% of their customers already use multiple products. DoorDash is behind on this metric but moving in the same direction, just they don’t have rides, so less diversity in verticals.
4/ Deliveroo, 5 months post-acquisition: growing faster than expected at the same EBITDA contribution, orders accelerating in Q4. Grocery/retail unit economics turning positive H2 2026.
5/ The replatform. DoorDash expanded internationally through acquisitions – Wolt in 2022, Deliveroo in 2025 – and inherited three separate tech stacks for essentially the same business. Tony called this “not ideal” – understatement of the call. Uber built internationally from one platform and never had this problem.
So the plan now is one global stack, mostly done in 2026. Early proof already: features from US or Wolt markets shipping into Deliveroo with immediate impact, Tony sees hundreds of those opportunities.
Parallel overhead bleeds into 2027, but once done, DD will have what Uber has had from the start – plus market leadership across Europe.