Something strange happens when you order a late-night burrito in a major American city.
Your driver may know more about your street than your neighbors do—and there's no oversight committee for that.
Over the past few years, DoorDash has built one of the largest real-time human sensor networks in the country.
Roughly two million drivers, spread across nearly every zip code, each carrying a GPS-pinging phone.
During the pandemic, the company leaned into it: DoorDash partnered with local governments to document which restaurants were closed, which streets were flooded, and where testing sites sat empty.
Now consider what that network actually sees.
A driver waits outside your building at 11 p.m.
Multiply that by millions of daily deliveries and you have a live map of American life—who's home, who's working, which blocks are thriving and which are hollowing out.
DoorDash says it uses this data to estimate wait times and improve routing.
But data has a way of outgrowing its original job description.
The company has already tested driver-based "food safety" reporting in some markets.
It has experimented with having dashers photograph storefronts.
And during the 2020 election, DoorDash offered drivers paid time off to work polling locations—a move praised as civic, but one that also demonstrated how quickly this fleet can pivot to tasks beyond food.
Then there's the surveillance question nobody in Silicon Valley wants to answer directly.
DoorDash has been accused in lawsuits of tracking drivers even after they clock out—an allegation the company denies, but one that has surfaced repeatedly in arbitration filings.
If the app can locate a driver at 2 a.m., it can locate anyone who accepts a delivery.
Combine that with the company's 2021 acquisition of food-tech startup Chowbotics and its push into grocery and convenience, and the data pie keeps growing.
DoorDash sits on a private, nationwide, on-demand observation grid—one that can be aimed at almost any problem, with drivers who are paid per task and bound by arbitration clauses.
The question is who gets to point it, and who gets watched while they do. **Closing take:** We handed a private company a real-time map of American neighborhoods because we wanted pad thai delivered in 20 minutes.
Final Thoughts
But if this fleet can report on closed restaurants and flooded streets, it can report on almost anything else—and we should be asking who's reading the reports before they knock.