🔍 Read the full analysis: The State Of Agent Shopping, Seen Through DoorDash’s Latest Numbers on ThorstenMeyerAI.com
Get smart everyday buys 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 have produced “maybe several hundred orders in total” since they began operating, describing the use as low. DoorDash handled 993 million orders in the third quarter of 2026, while Meta, Amazon and Visa have also described agent-driven shopping as early or limited. The figures measure different things, and Meta’s report of shopping activity on Muse does not establish comparable completed-order volumes.
DoorDash CEO Tony Xu said third-party AI agents have generated “maybe several hundred orders in total” since they began operating, a small volume beside the company’s 993 million orders in the third quarter of 2026. The comment, made on DoorDash’s October 29 earnings call, offers a concrete measure of how little outside agents currently contribute to ordering on a major delivery platform, even as technology companies promote AI as a new way to shop.
Xu made the remark in response to an analyst asking whether DoorDash planned to serve food-ordering agents itself or leave that work to broader assistants such as Meta’s Muse. He said the area remained “very dynamic,” but cautioned against overstating adoption. For a straightforward task such as ordering food, he said, the predominant use case is still DoorDash’s own app.
Xu returned to the figure later in the call, saying agents had generated several hundred orders “across the time that they’ve been up and running.” He called it a “fairly low use case”. Asked whether agents could affect advertising or share in advertising revenue, he said opportunities might emerge, but low current volumes meant there was not yet corresponding value.
DoorDash reported 993 million total orders in the third quarter, up 28% from a year earlier. The company’s order total covers the quarter; Xu’s rounded agent figure covers the entire period those agents have been active. The two periods are not directly matched. As an illustration only, if “several hundred” meant 1,000 orders and all occurred during the quarter, that would be about one in a million of DoorDash’s quarterly orders. That is a rough comparison, not a company-reported rate.
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
Agent Orders Remain Marginal at DoorDash
The disclosure grounds a broad debate about AI shopping in a real-world transaction count. Agents may help people find products or make recommendations, but DoorDash’s remarks indicate that outside agents have not yet become a meaningful ordering channel on its platform. That matters to retailers and delivery businesses weighing how much to invest in agent access, checkout systems and partnerships.
It also bears on advertising. If shoppers start product searches and purchases inside an assistant, retailers could lose influence over the discovery process and the ads attached to it. Xu said DoorDash sees possible advertising opportunities within agents, but described current volumes as too low for that to have value yet. His comments identify a potential business issue, not evidence that agent shopping has already changed DoorDash’s advertising results.
For consumers, the immediate takeaway is narrower than claims that AI will soon handle routine purchases: at least on DoorDash, the company says people still mainly use its own app for food orders. Whether agents become more useful may depend on how smoothly they can handle payment, preferences, substitutions and delivery problems, not just on their ability to start a search.
Products Worth Considering
Apple iMac 2024 All-in-One Desktop with M4 Chip, 24‑inch Retina Display, 16GB Unified Memory, 256GB SSD, Silver
Dell 24 Plus Monitor S2425HSM – 23.8 FHD (1920×1080) 144Hz 1ms, 2 x 3W Speakers, HDMI, Height/Tilt/Pivot/Swivel Adjustability, AMD FreeSync
Dell XPS 13 9350 AI Professional Laptop 13.4 120Hz FHD+ Display
As an affiliate, we earn on qualifying purchases.
Other Companies Describe Early Adoption
DoorDash’s remarks came amid a cluster of earnings-call comments about AI shopping. 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 roughly six weeks, and that shopping was its most common use. Meta also said more than a quarter of users engaged in shopping and reported an average purchase size above $200. Those are Meta’s own figures; the source material does not provide a total sales value or an independently verified count of completed purchases comparable to DoorDash’s order tally.
On October 27, Visa CEO Ryan McInerney described an “albeit limited number of agentic transactions” and said trust was “the biggest barrier to scale.” Amazon CEO Andy Jassy, speaking October 29, said few customers were doing much shopping through outside, general-purpose agents because the experience was not yet as convenient as buying directly from a retailer. These comments broadly point to early use, but each company is discussing a different measure: user set-ups, transactions or customer behavior.
DoorDash has also framed its technology strategy around what Xu has called “atoms vs. bits”: software and AI on one side, and the physical work of storing, preparing and delivering goods on the other. The company has discussed its own agents and investments in physical infrastructure. The source material also describes a planned $17 billion deal to buy Instacart; its status and terms are not detailed in the earnings-call excerpts provided here.
““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
What the Order Count Cannot Show
DoorDash did not give an exact count, a start date for the measurement period, or a breakdown by agent. Xu’s description, “maybe several hundred,” is approximate, and the period covers the time the agents have been running rather than a specified quarter. The comparison with 993 million quarterly orders is therefore illustrative only.
The disclosure also does not show how many people tried to use an agent, how often attempts failed, or whether customers completed purchases through other arrangements that DoorDash would not classify as third-party-agent orders. Meta’s Muse metrics are not directly comparable: user set-ups and shopping engagement are different from a count of completed orders on a delivery platform. The source material does not establish whether the companies use consistent definitions of an agentic transaction.
It remains unclear which technical or commercial barriers account for the low DoorDash volume, how the figure may change, and whether agent-originated orders could eventually support advertising or other revenue. Xu’s comments do not amount to a forecast of future adoption.
Watch for Volume and Checkout Changes
The next useful signal will be whether DoorDash or other platforms publish comparable counts over a defined period, including an exact measure of completed orders and the agents involved. Future earnings calls may also clarify whether companies can distinguish transactions initiated by an assistant from purchases made after a customer moves to a retailer’s app or website.
For DoorDash, the company’s comments leave open whether it will deepen integrations with outside agents, build more ordering tools of its own, or emphasize its delivery network. Xu said potential advertising opportunities inside agents could become relevant if volume grows, but said current use is too low to create that value. No specific launch, timetable or adoption target was given in the material.
For now, the confirmed point is a modest one: DoorDash says third-party agents have generated several hundred orders over their operating period, while its platform handled 993 million orders in one quarter. The scale and direction of any future change remain to be measured.
Key Questions
How many orders have third-party AI agents generated on DoorDash?
CEO Tony Xu said agents had generated “maybe several hundred orders in total” since they began operating. He did not give an exact count or a precise start date.
How does that compare with DoorDash’s overall order volume?
DoorDash reported 993 million orders in the third quarter of 2026, up 28% year over year. The agent figure covers the agents’ operating period, not necessarily that quarter, so any ratio is only an illustration.
Does Meta report the same kind of shopping figure?
No. Meta reported Muse set-ups, shopping engagement and an average purchase size above $200. Those measures are not directly comparable with DoorDash’s reported count of orders generated by outside agents.
Why might agent shopping still be limited?
The executives cited in the source material point to convenience and trust. Amazon’s Andy Jassy said outside agents were not yet more convenient than shopping directly, while Visa’s Ryan McInerney called trust a barrier to scale. The material does not establish a single cause for DoorDash’s low order count.
Has DoorDash said when agent orders will grow?
No adoption target or timetable was provided. Xu described the area as dynamic and said advertising opportunities could emerge, but that current volumes are too low to create value from them yet.
Source: ThorstenMeyerAI.com
