The Verge has published an analysis arguing that artificial intelligence could fundamentally upend food delivery — a business built on being the software layer between hungry customers and restaurants.

The report centers on a structural question for the restaurant industry: if AI agents start placing orders on behalf of consumers, what happens to the apps that currently own that customer relationship?

Why does this matter for operators?

Third-party delivery platforms — DoorDash, Uber Eats and Grubhub in the U.S. — sit between restaurants and diners, and they extract commissions that operators have long complained compress already-thin margins. The Verge's piece frames agentic AI as a potential bypass of that intermediary position.

The logic runs as follows. Today, a customer opens a delivery app, browses menus, compares prices and places an order; the platform takes its cut and controls the data. Tomorrow, according to the analysis, a consumer could simply tell an AI assistant what they want to eat, and the agent could search menus, compare total costs including fees, and place the order wherever the economics are best — directly with the restaurant if that route is cheapest.

That scenario, if it materializes, would shift leverage back toward operators that maintain their own ordering channels.

What does agentic AI actually replace?

The technology at issue is not kitchen automation or robotics. It is the discovery-and-transaction layer — search rankings, menu browsing, promo surfaces and checkout flows that platforms currently monetize.

For restaurants, that layer is expensive. It is the source of commission structures, advertising pay-to-play placement fees, and the customer data that operators surrender when orders arrive through a marketplace rather than a first-party channel. An AI agent working for the diner rather than the marketplace could, in theory, ignore the ad placements and optimize purely on price, speed and quality signals.

The Verge's framing implies the platforms' most defensible assets would then be their logistics networks — the driver fleets — rather than their storefront apps.

Who pays and who gains?

The distributional question cuts both ways. If AI agents route orders around marketplaces, operators could recapture commission dollars on transactions that originate through agent-driven discovery. But the same dynamic could intensify price competition, because an agent comparing across venues makes total delivered cost far more transparent than a human scrolling one app at a time.

For the delivery platforms themselves, the analysis positions them as the most exposed party — intermediaries whose value depends on customers not having a cheaper, smarter way to find and order food.

How immediate is the threat?

The Verge presents this as a plausible trajectory rather than a present-day disruption. Consumer-facing AI agents capable of reliably completing restaurant transactions — handling payment, substitutions, timing and error resolution — remain early in their development.

Still, the direction matters for capital planning. Operators weighing investments in first-party ordering infrastructure, direct-delivery pilots and menu-data standardization may find the calculus shifting if agent-mediated ordering gains share. Cleaner, machine-readable menu data becomes an asset when the customer is an AI.

The restaurant industry has already lived through one disintermediation scare — the rise of the marketplaces themselves a decade ago, which pulled ordering volume away from operator-owned channels. The Verge's analysis suggests the next shift could run the other way, and the platforms, not the restaurants, may be the ones scrambling to defend their position.