Amazon Drops Grubhub for Uber Eats: What Marketplace Supply-Side Replacement Looks Like at Scale

Amazon has ended its food delivery partnership with Grubhub and replaced it with Uber Eats as the official restaurant delivery partner for Prime members in the US. Prime members can now order food directly through Amazon.com without creating an Uber account, with free delivery on orders over $25. The Grubhub deal, originally struck in 2022, is effectively terminated — though existing Grubhub+ activations run through December 2027. Amazon and Uber already had an extensive existing relationship across cloud, robotaxis, and Alexa integration.

This is not simply a vendor swap. It signals that Amazon is treating food delivery as a retention lever for Prime — not a standalone business. The decision to cut Grubhub came down to one criterion: which supply-side partner delivers the best demand-side experience for Prime members. Grubhub's market share decline and ownership instability (sold from Just Eat to Wonder) made it a liability. Uber Eats, with deeper liquidity and broader restaurant coverage, is a stronger supply anchor. The deeper signal: in marketplace ecosystems, your embedded partners are only as valuable as the liquidity and reliability they bring to your users — a principle central to how you build a successful marketplace.

SUPPLY: Uber Eats brings significantly deeper restaurant supply than Grubhub — more cities, more cuisines, higher fill rates. When Amazon evaluates a supply partner, breadth and reliability of that supply directly affects the perceived value of Prime. Thin supply kills embedded marketplace features.

DEMAND: Amazon is using food delivery as a demand retention tool, not a revenue line. Prime members who get free delivery through Uber Eats have another reason not to cancel their membership. This is demand-side lock-in through supply-side access.

LIQUIDITY: Grubhub's shrinking market share meant lower liquidity — fewer orders matched to fewer drivers to fewer restaurants. Uber Eats has stronger liquidity density in most US metros, which translates to better fulfillment rates and fewer failed transactions.

TRUST: Amazon removed account creation friction entirely. Prime members don't need an Uber account. This is a trust and onboarding decision — reducing the number of steps between intent and transaction eliminates drop-off and builds trust in the embedded experience.

GROWTH: The partnership deepens Amazon's ecosystem lock-in. Every new use case added to Prime raises the switching cost for members. For Uber Eats, access to Amazon's 200M+ Prime members is a customer acquisition channel at near-zero marginal cost.

ONBOARDING: The no-account-required flow is a masterclass in onboarding design. The biggest barrier to trying a new marketplace is registration. Eliminating it entirely for an already-trusted user base (Prime members) dramatically compresses the funnel.

MONETIZATION: Amazon is not monetizing the food delivery transaction directly. It monetizes through Prime retention. Uber Eats monetizes through order volume and restaurant commissions. Each party extracts value from a different layer — this is what a healthy embedded marketplace partnership looks like, and it reflects many of the marketplace launch best practices that separate sustainable models from ones that collapse under their own complexity.

• Audit every supply-side partner you embed in your marketplace by one metric: does their liquidity hold up when your demand volume peaks? If they can't fulfill consistently, they're hurting your brand, not helping it.

• If you're building an ecosystem or membership product, treat supply-side access as a retention mechanism — not just a feature. Each additional use case raises switching costs for your users.

• Eliminate account creation wherever possible when embedding third-party supply. The biggest onboarding drop-off is registration. Negotiate single-sign-on or no-account-required flows into your partnership terms upfront.

• Monitor your supply partners' market health — ownership changes, market share shifts, and geographic contraction are early warning signs that their liquidity is thinning. Build exit clauses into contracts.

• When evaluating supply partners, weight their fulfillment reliability and geographic coverage over their brand recognition. A well-known but thinly liquid partner will underdeliver your users more than a lesser-known but operationally strong one.

• Structure partnerships so both sides extract value from different layers. Amazon takes retention value; Uber Eats takes order volume. Misaligned value extraction leads to renegotiation friction and eventual collapse.