How gig delivery apps became retail infrastructure

When Home Depot promises to bring tools to a job site in an hour and Dollar Tree drops party supplies at a shopper's doorstep, neither is moving the goods. That work belongs to Instacart, DoorDash, or Uber Eats, three companies that connect stores to drivers so their partners can offer same-day delivery.

Tony Xu completed DoorDash's first delivery more than 12 years ago as a school project he named PaloAltoDelivery.com, handing over an order of chicken pad thai with spring rolls and keeping the fee himself. The company he built from that now averages more than seven million deliveries a day, and Xu has been explicit that restaurants were always meant to be the opening act. "There is a reason why the name of the company is DoorDash, unrestricted to food," he wrote in his 2024 letter to shareholders.

"That's because local economies do not grow from restaurants alone, and it was a matter of time before we'd serve other types of retailers."

Instacart, DoorDash, and Uber Eats each pushed past the niche they started in until their network carried almost any store a shopper might want. DoorDash, which launched grocery and convenience in mid-2020 and added home improvement in 2024, now serves 44 of the top 100 U.S. retailers alongside its restaurants.

Uber took the same path from the other direction, recasting grocery, alcohol, convenience, and general merchandise as a single "Grocery & Retail" business. It added Home Depot for contractors ordering supplies to job sites and brought almost 9,000 Dollar Tree stores onboard in an August 2025 partnership. Instacart, which began in groceries, now connects more than 2,200 retail banners representing almost 100,000 stores, a network reaching more than 98% of North American households.

While each company started out differently, they together became the fulfillment backbone for thousands of stores that would rather pay for their services than build a shipping network of their own.

Instead of building new fulfillment centers as Amazon did, these companies make same-day delivery work by turning the stores themselves into shipping warehouses. The inventory is already sitting on shelves close to the buyer, so a chain can dispatch a courier to move product off those shelves rather than choose between building its own logistics or ceding fast delivery to Amazon.

Instacart says more than half its orders go to shoppers who are already standing in or within a mile of the store, close enough to make delivery in an under an hour feasible without incurring high mileage. Gig workers are all too happy to take these orders, especially at high volumes. Americans placed $9.6 billion in online grocery orders in November 2024, up 17.8% from a year earlier, according to the Brick Meets Click/Mercatus survey, with delivery sales alone climbing 22% to $3.9 billion.

For years, the companies moving those orders lost money doing it. That changed in 2024, when DoorDash turned its first annual profit under generally accepted accounting principles — $123 million on $10.7 billion in revenue, according to its fourth-quarter results — and Instacart cleared $457 million in net income, its own filings show. Only with that turn did a retailer's reliance on these networks stop being a bet on partners that might not survive.

Now that these delivery services can pay their own way, they're wringing more out of the retailers riding on them. DoorDash charges restaurants commissions of up to 30% of each order's subtotal on its Premier plan, according to its merchant pricing, and after payment processing and all-but-required advertising get added in, Uber Eats' blended cost commonly runs 25% to 35%, according to an analysis by direct-ordering vendor Zay-OS.

On a $30 order at 25%, one in four dollars leaves the restaurant. A busy independent store can hand third-party platforms $48,000 or more a year.

The platforms increasingly control not just the fee but the customer and the price at checkout. A joint investigation by the Groundwork Collaborative, Consumer Reports, and More Perfect Union found that Instacart, using pricing software called Eversight, ran hidden randomized experiments that could add as much as 23% to the cost of an identical item ordered from the same store at the same moment. Shoppers at a Safeway in Washington, D.C., found a dozen Lucerne eggs carrying five prices at once, from $3.99 to $4.79.

Lindsay Owens, who heads the Groundwork Collaborative, called it "a high-tech game of pricing roulette," estimating the average basket swing of about 7% could cost a family of four roughly $1,200 a year. Instacart said in a July 2026 post that a few of its retail partners had wound down the tests and that the experiments never used customer data.

The people who actually carry the goods sit at the center of a fight over whether they're employees owed wages and benefits or independent contractors who are not. The California Supreme Court left that question largely settled in the platforms' favor in July 2024, when it upheld Proposition 22 in Castellanos v. State of California, ruling that the measure voters passed with 58.6% support doesn't conflict with the state constitution and can keep app-based drivers classified as contractors outside the reach of workers' compensation laws.

New York City regulators are raising what the platforms must pay those workers anyway, and each expansion of that obligation eats into the companies' profits. The city's Department of Consumer and Worker Protection first required restaurant apps to pay a delivery-worker minimum, and in a December 2025 final rule, the department pulled grocery apps under the same floor. It has kept lifting the number to track the cost of living, setting the rate at $22.13 for the first pay period starting on or after April 1, 2026 after applying a 3.2% inflation adjustment. The department plans to cover all delivery apps in early 2027, meaning still more of the platforms' business will fall under a wage obligation they once avoided.

Put together, these pressures point past the arrangement itself to who ends up holding the leverage. With commissions climbing, pricing software rewriting shelf tags, and the wage floor rising under couriers the retailer doesn't employ, the stores lose more and more control of the products they carry, and often can't even see what the total comes out to.

The retailer rents the network it could never build, but from a landlord now profitable enough to keep raising the terms. The more indispensable the backbone becomes, the less the store on the label decides how its own sale is run.