🔍 Read the full analysis: What “Several Hundred” Orders Mean For AI-Powered Shopping on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
DoorDash CEO Tony Xu said third-party AI agents had generated “maybe several hundred orders in total” since they began operating, while DoorDash handled 993 million orders in the third quarter of 2026. Amazon and Visa executives also described agent-led shopping as limited so far, even as Meta says shopping is the most common use of its Muse agent.
DoorDash CEO Tony Xu said third-party AI agents have generated “maybe several hundred orders in total” since they began operating, a small early volume compared with the 993 million orders the delivery platform processed in the third quarter of 2026. His comments, made on the company’s October 29 earnings call, offer a concrete measure of how little outside agents are currently contributing to orders at a major commerce platform.
Xu gave the figure in response to an analyst asking whether DoorDash wanted to power food-ordering agents itself or leave that work to general-purpose services such as Meta’s Muse. He described the area as “still a very dynamic” one, but said that for an easy task such as ordering food, the predominant use case remains DoorDash’s own app. Xu later repeated that agents had generated several hundred orders across the time they had been available and called it “a fairly low use case.”
DoorDash’s third-quarter results put the scale in perspective: the company reported 993 million total orders, up 28% from a year earlier. Xu did not give an exact agent-order count or specify a start date for the period covered by “in total.” If “several hundred” is generously treated as 1,000, and all those orders are assumed to have occurred during the quarter, that would be roughly one order per million DoorDash orders. That is an illustrative calculation, not a company-reported share, and the period mismatch means it should not be read as a precise rate.
Xu also said there was not yet enough volume to create meaningful advertising value within third-party agents. He said advertising opportunities could develop inside those services, but “right now, with the low volumes,” there “isn’t yet that value.” The remarks came as analysts asked how agent-based shopping might affect DoorDash’s business and its relationship with outside platforms.
Infographic 01 · The number
993 million orders. “Several hundred” from AI agents.
What DoorDash’s CEO said about orders placed by outside AI agents, next to the company’s own order count.
Drawn to scale
Q3 2026 orders vs. a generous upper bound for agent orders
The second bar is not missing. At this scale it is about one ten-thousandth of one percent wide.
“You never want to fight gravity. But at the same time, you don’t want to be delusional … we’re seeing maybe several hundred orders in total from some of these third-party agents.”Tony Xu, CEO of DoorDash · Q3 2026 earnings call, 29 Oct 2026
Sources: DoorDash Q3 2026 results (29 Oct 2026); DoorDash earnings call (29 Oct). Ratio and per-day figure: my calculation. “Several hundred” is a spoken, rounded figure.
Powered by Thorsten Meyer AI
Why Agent Orders Remain Small
The figure complicates broad expectations that AI assistants are already shifting everyday purchases away from retailer and delivery apps. At DoorDash, the disclosed count suggests that outside agents have not yet become a meaningful order channel, at least by the measure Xu described. For customers, the practical point is that using an AI assistant to find or discuss a purchase does not necessarily mean it completes the transaction.
The gap also matters to businesses deciding where to invest. If customers continue to place most orders in a company’s own app, the platform retains control over checkout, customer relationships and advertising. Xu’s comments indicate that DoorDash sees possible future value in agent advertising but does not yet see enough third-party agent activity to support it. The order count is a snapshot of current use, not a forecast of what the channel will become.
Executives at other companies described similar limits during the same week. Amazon CEO Andy Jassy said few customers were doing much shopping through outside general-purpose agents, adding that the experience was not yet better or more convenient than shopping directly with a retailer. Visa CEO Ryan McInerney called trust “the biggest barrier to scale” and described agent transactions as limited. Those statements point to adoption and payment confidence as unresolved challenges, rather than evidence that agent commerce has failed.
Products Worth Considering
Anker SOLIX S2000 Portable Power Station, 2,010Wh, 1,500W
Bose SoundLink Flex (2nd Gen) Portable Bluetooth Speaker
Jackery Explorer 300 Portable Power Station, 292Wh
As an affiliate, we earn on qualifying purchases.
What Companies Reported This Week
The comments came amid rising attention to AI shopping assistants. Meta CEO Mark Zuckerberg said on October 21 that more than 10 million people in the United States and Canada had set up Muse in about six weeks. He said shopping was its most common use case, with more than a quarter of users engaging in shopping and an average purchase size above $200. Those are Meta’s figures; the source material provides no independent audit or total sales value, and the figures do not establish how often people complete ordinary purchases such as dinner orders.
Visa’s McInerney discussed agent-led transactions on October 27. On October 29, the same day as DoorDash’s call, Amazon’s Jassy characterized shopping through outside agents as limited and said the experience was not yet more convenient than researching or ordering directly from a retailer. These comments are not directly comparable: the companies described different products, activities and measures. Meta’s user and purchase claims do not measure DoorDash orders, while DoorDash’s figure concerns orders attributed to third-party agents across an unspecified operating period.
DoorDash’s own business is built around an app and a delivery network, so the distinction between recommending a purchase and completing it matters. Xu framed the company’s technology approach as a balance between software and physical operations, referring to “atoms vs bits.” The source material also describes DoorDash’s investment in its physical network, including warehouses and autonomous delivery, alongside its development of its own agents. That strategy reflects the company’s existing delivery business, not proof that outside agents will or will not gain traction.
““We’re seeing maybe several hundred orders in total from some of these third-party agents.””
— Tony Xu, DoorDash CEO, on the October 29 earnings call
Limits of the Order Count
DoorDash did not define exactly what “several hundred” means, state the precise number of orders, or give the period over which it counted them. Xu said the figure covered the time the third-party agents had been running, which is not necessarily the same span as the third quarter. The comparison with 993 million quarterly orders is therefore illustrative only, and the exact share of orders placed by agents cannot be calculated from the disclosed figures.
The comments also do not establish whether the count includes every kind of third-party agent, how DoorDash identifies an order as agent-generated, or whether customers used an agent before completing checkout through DoorDash. The source material does not provide a transaction breakdown by agent or details on repeat use. Meta’s separate Muse figures likewise do not show how much shopping value was completed or how those purchases compare with routine delivery orders.
It remains unclear whether low volume reflects limited consumer interest, product or payment friction, restricted access to retailer systems, or some combination. Executives cited convenience and trust concerns, but did not provide data that would isolate their effects. The figures describe activity at this early stage; they do not establish how adoption will change.
Signals to Watch in Upcoming Results
Future earnings calls and company disclosures may show whether third-party agents begin contributing a larger number of completed orders. Useful details would include a defined measurement period, an exact count, the share of total transactions, and whether customers completed checkout within the agent or on the retailer’s own service. Without those details, headline counts from different companies will remain difficult to compare.
Investors and customers will also be watching whether platforms make agent checkout more convenient and whether payment providers can address concerns about authorizing purchases. DoorDash has said it sees possible advertising opportunities in agents but that current volumes are too low to make those opportunities valuable. For now, the next test is whether companies report sustained growth in completed transactions, rather than user sign-ups or shopping-related conversations alone.
Key Questions
How many orders did DoorDash say AI agents placed?
CEO Tony Xu said third-party agents had generated “maybe several hundred orders in total.” He did not give an exact count or specify the full period covered.
How does that compare with DoorDash’s total orders?
DoorDash reported 993 million orders in the third quarter of 2026. Treating “several hundred” as 1,000 and assuming all occurred during that quarter gives roughly one agent order per million total orders, but that is an illustrative calculation because the periods may differ.
Did DoorDash say customers are not using AI for shopping?
No. Xu’s comments covered orders from third-party agents at DoorDash. They do not measure every kind of shopping activity, and Meta separately reported that shopping was the most common use of its Muse agent.
Why might customers still order directly through apps?
Xu said DoorDash’s app remains the predominant way to order food. Amazon’s CEO said outside-agent shopping was not yet more convenient than ordering directly, while Visa’s CEO cited trust as a barrier. The comments identify possible obstacles but do not measure their individual effects.
Source: ThorstenMeyerAI.com
