Source time Edition date
IndustryArs Technica AI
Basically no one is using AI agents for their DoorDash orders
几乎没人用AI代理下DoorDash订单
Worth reading because it shows adoption limits: DoorDash's new e-commerce agent use case remains mostly unused, so the question is who decides whether to keep investing in it.
Judgment
DoorDash CEO Tony Xu must decide whether to keep funding third-party AI-agent ordering and the MCP beta after only "several hundred orders" against just shy of 1 billion Q3 orders, or redirect that budget to internal Ask DoorDash, which shows stronger basket economics.
Basis
On DoorDash’s Q3 2026 earnings call, CEO Tony Xu said third-party AI agents generated “several hundred orders” since launch and called it a “very low use case” for a company that handled just shy of 1 billion orders total in Q3 2026. The mechanism is partly structural: DoorDash’s MCP connector was less than a month old, open only to a small number of early beta partners, and mainly used for corporate lunch orders; the third-party agent figure also excludes internal Ask DoorDash. DoorDash said Ask DoorDash baskets are built “five times faster,” and Ask DoorDash orders “average nearly 50% higher basket value and about 60% more unique items than the same consumers’ traditional orders,” with nearly half going to local spots the consumer has never tried. Xu added: “you never want to fight gravity,” but “you don’t want to be delusional.”
Unclear The source does not say what DoorDash spends on the MCP connector, how many early beta partners are live, or how third-party agent order volume has changed month over month.
Next On the next earnings call, check whether DoorDash discloses third-party agent order volume or MCP partner count, and whether it updates Ask DoorDash basket and discovery statistics.