The Maze: Food delivery is not really a fight over meals. It is a fight over the recurring order. theDelivery.World's June 24 issue shows why: Swiggy / Instamart's India gross transaction value moved from 61% restaurant and 39% quick commerce to 53% and 47% in one year, while Delivery Hero says grocery is now about 15% of group value. The kitchen still cooks. The platform increasingly owns the habit.
Restaurant delivery is becoming an order business with meals attached. The source calls this shift "the breakout": platforms that built frequency in restaurant delivery are pushing the same customer base, riders and app habits into grocery, convenience and pharmacy. Swiggy is the cleanest case. In India, Instamart climbed to 47% of gross transaction value in Q1 2026, nearly matching food delivery's 53%. That is not a side quest. It is a second use case being loaded onto the same demand loop.
The model works because the order repeats and the meal does not. A restaurant sells one dinner. A platform owns the app, address, payment card, search history and next trigger. That difference is why the article argues commission rates explain less than operators think. Two restaurants can pay the same platform fee, but the one whose customer returns through the app is effectively paying a recurring access charge on its own demand. The value sits with the company that controls the return path.
Grocery makes the platform relationship more useful and more expensive to lose. Delivery Hero's grocery share is about 15%, with a caveat: the remaining 85% is restaurant plus other verticals, not pure restaurant. Talabat's MENA split is roughly three quarters restaurant and one quarter grocery / quick commerce. Instacart shows the reverse path: a grocery-first platform adding restaurants. The direction is the point. A customer who orders dinner tonight can be pointed to milk tomorrow morning.
The economics are still uneven. The article is careful here. Grocery and quick commerce are mostly a volume story, not yet the profit engine. Restaurant delivery still funds much of the expansion. Blinkit and Swiggy have built large dark-store networks, but those warehouses are bets on recurring demand, not proof that every basket is profitable. That matters because operators can mistake volume migration for margin migration. The customer relationship is the asset. Profit takes longer.
AI agents could reopen the whole ownership question. The order already moved once, from the restaurant phone line to the platform tap. If routine purchases move into assistants, the decision may sit with the agent owner instead of DoorDash, Swiggy, talabat or Blinkit. The source frames this cautiously. Timing is uncertain. But the logic is consistent: value follows the recurring decision, wherever it lands.
Why it matters: Restaurants have spent years arguing about platform commissions. That is the visible bill. The deeper issue is customer ownership. If the platform owns the recurring order, it can sell ads, subscriptions, grocery and convenience against the same habit. If an AI assistant owns the next order, even the platform may become a supplier. The winner is not always who cooks, delivers or lists the product. It is who captures the next default decision.
