Building an Autonomous Delivery Experience with DoorDash Co-Founders Andy Fang and Stanley Tang

DoorDash co-founders Andy Fang and Stanley Tang set out the company’s push into “agentic commerce” and autonomous delivery, presenting both as extensions of a large-scale local commerce network. Fang described “Ask DoorDash”, a natural-language interface for restaurant and grocery ordering, and claimed it is shifting behaviour: he said half of restaurant trajectories using the tool end in orders from restaurants a customer has not used before, and that grocery baskets are around 40% larger.

Tang traced DoorDash’s autonomy work back to 2018, initially pursued through partnerships rather than building hardware. He argued those collaborations validated autonomy as inevitable, but also exposed gaps in existing robots: sidewalk units were too slow for typical 3–5 mile deliveries, while robotaxis are oversized and poorly suited to pick-up/drop-off constraints. Tang said this led DoorDash to build DOT in-house — a roughly 300lb, 20mph delivery robot operating in Phoenix — and to create an “autonomous delivery platform” of APIs and operational tooling around it.

The conversation also addressed scaling constraints beyond autonomy software, including fleet operations, mapping the “first and last 100 feet” using DoorDash’s historical drop-off data, and hardware manufacturing. Tang said DoorDash partnered with micromobility firm Also, spun out of Rivian, to help scale vehicle production.

On internal AI adoption, Fang said the company’s AI spend rose about 20x from January to June before flattening, and described efforts to benchmark model performance and ROI via a coding-focused evaluation called Dashbench.

• DoorDash co-founder claims Ask DoorDash users order from new restaurants in 50% of trajectories.
• DoorDash co-founder claims grocery baskets are about 40% larger when using the natural-language tool.
• Stanley Tang says DoorDash began autonomy and robotics work in 2018 through early experiments and partnerships.
• Tang describes DOT as an in-house autonomous delivery robot (about 300lb, up to 20mph) operating in Phoenix.
• Tang says scaling autonomy is increasingly constrained by operations and hardware manufacturing, not just the driving stack.
• Fang says DoorDash’s AI spend increased roughly 20x from January to June and has since flatlined as ROI work ramps up.
• Tang predicts DoorDash will have more Dashers in 10 years even with widespread robots and drones.