DoorDash earned $244 million in Q3 2025, up from $162 million a year earlier. Orders rose 21% to 776 million. The uncomfortable part of its shareholder letter was the plan to invest several hundred million dollars more in 2026 than in 2025. I originally wrote that as the total budget and treated every dollar as an autonomous-delivery bet. Both readings were wrong.
I still think spending on a common technology platform makes sense. DoorDash had added Wolt and Deliveroo; keeping separate systems for product development and operations would make every launch slower. But “global stack” can cover a lot of expensive engineering work. The company should show whether a shared platform helps a local team launch a feature faster or run orders more cheaply.
DoorDash reported a 13.8% net revenue margin, up from 13.5% a year earlier. It attributed the change partly to advertising, fewer credits and refunds, and lower Dasher costs relative to order value. That is the scorecard I would place beside the spending plan. If order economics improve while the company integrates acquisitions, the investment has an argument. If costs rise and the margin stops moving, management owes a better explanation.
I also want separate evidence for Serve’s robots, DoorDash’s Dot, and the Waymo Phoenix test. Robot cost claims mean little without completed orders, handoffs, intervention work, and the share of deliveries these vehicles can actually serve. No public number in this earnings release says robots beat couriers on a completed-order basis.
The stock reaction was a judgment on future spending, not proof that the spending was good or bad. I would rather check what the platform can do with each extra dollar. The three-company comparison makes the same point from a different angle: demand was easy to find; cheaper service was not.