Apple Wants to Price the App Store’s Reach, Not the Cost of a Link
Executive summary
Apple’s latest proposal in the Epic Games litigation asks a federal court to permit commissions of 15%, 10%, or 5% on purchases completed outside the App Store after a user follows an in-app link. The filing is important because Apple also says that a fee limited to the Ninth Circuit’s narrow definition of costs necessary to coordinate external links would be essentially zero.
The mainstream reading is that external checkout has opened the U.S. iOS market and that the remaining dispute is simply the size of a permissible fee. That misses the larger contest. Apple is trying to move the economic basis of the toll from payment processing to the value of distribution, tools, intellectual property, and access to users. If that theory prevails, web checkout may lower the platform take but will not automatically create direct-to-consumer economics.
Our contrarian conclusion is conditional: the most durable platform rent may be an attribution claim rather than a payment fee. The court may reject Apple’s proposed rates, and Epic has not yet filed its full response. But operators should plan for a future in which controlling billing and customer data does not mean escaping the acquisition channel’s claim on revenue.
Market context
The dispute began with anti-steering rules, but it has become a test of what a platform is allowed to monetize after the transaction leaves its rails. The original injunction prevented Apple from barring developers from using buttons, external links, or other calls to action that direct users to purchasing mechanisms outside the App Store.
Apple’s first U.S. implementation charged a commission of 27% on linked-out purchases for standard apps, only three percentage points below the typical 30% in-app purchase commission, and imposed restrictions on link design. In its December 2025 opinion, the Ninth Circuit upheld the finding that the 27% rate had a prohibitive effect. It also concluded that Apple’s restrictions made links difficult to use and affirmed the core contempt findings.
The appeals court did, however, send the commission question back for further work. Its direction left room for Apple to recover costs genuinely and reasonably necessary to coordinate external links and left the parties to litigate the treatment of intellectual property. Apple’s August 13 remand filing is its proposed answer.
Apple proposes three percentage rates: 15% for standard apps that ordinarily face a 30% in-app purchase commission; 10% for apps in the Video Partner, News Partner, and Mini Apps Partner programs and for subscription renewals; and 5% for Small Business Program apps. The Verge’s review of the filing reports that Apple describes costs under the court’s narrow necessary-cost standard as effectively zero while arguing that the broader economic value of App Store distribution and technology deserves compensation.
That distinction is the case.
Findings
Finding 1: The proposed fee is a claim on distribution value, not reimbursement for link-out work
If the economic question were limited to the cost of enabling a link, a percentage of transaction value would be hard to explain. The engineering, review, and administrative effort required to permit external links does not rise proportionally when a user buys a $100 annual subscription instead of a $10 monthly product.
Apple’s proposal therefore rests on a different theory. The developer acquired the customer through an app distributed, reviewed, and supported by Apple’s ecosystem; Apple argues that the subsequent web purchase still consumes value created by that ecosystem. The fee is presented as compensation for intellectual property and platform services, not merely payment handling.
That is economically coherent even if the proposed rates prove legally unacceptable. Marketplaces, app stores, affiliates, and retail-media networks routinely claim credit for a transaction that closes somewhere else. They use attribution windows, referral identifiers, or contractual reporting to connect discovery with conversion. Apple’s argument would turn an in-app link into the attribution event.
Finding 2: “External checkout” and “direct customer” are no longer synonyms
Moving payment to the web can still matter. Developers may gain control over checkout design, accepted payment methods, tax tooling, cancellation flows, offers, and first-party billing data. They may also pay a lower platform rate than under in-app purchase rules.
But those gains do not establish a fee-free direct relationship. A developer could operate the checkout and merchant account while owing Apple a percentage because the user arrived through the iOS app. Payment control, customer-data control, and distribution economics are separate layers.
For a standard app, Apple’s proposed 15% external-purchase fee would halve the nominal 30% platform commission before the developer pays its own processing, fraud, tax, support, and reconciliation costs. For a Small Business Program app, a 5% external fee would be ten percentage points below the typical 15% in-app rate. Those are potentially material improvements, but neither is equivalent to zero-cost steering.
The effect is especially important for subscriptions. Apple proposes 10% for subscription renewals. A renewal fee preserves a claim on revenue after the initial conversion, even though the ongoing billing relationship, support burden, and payment risk may sit with the developer. Subscription operators should evaluate the full retained lifetime value, not the first checkout alone.
Finding 3: The rate schedule would encode platform policy into business-model margins
Apple’s proposed tiers are not based only on transaction mechanics. They reflect existing program categories and developer status. That means two businesses using similar web-payment technology could face different external-purchase rates because Apple has classified them differently inside the App Store.
| App or purchase category | Apple’s proposed external-purchase fee | Stated comparison in Apple’s filing | Immediate operator question |
|---|---|---|---|
| Standard apps | 15% | Apps ordinarily subject to a 30% in-app purchase commission | Does the 15-point reduction cover processing, tax, support, and migration cost? |
| Video, News, and Mini Apps partner programs | 10% | Existing partner-program treatment | Is program eligibility durable, and can Apple revise it? |
| Subscription renewals | 10% | Renewals after the initial subscription transaction | How long does attribution persist, and what evidence supports it? |
| Small Business Program apps | 5% | Apps in Apple’s reduced-rate small-developer program | What happens when the developer crosses the program threshold? |
Source and methodology: Apple’s August 13, 2026 remand proffer, as reproduced by DocumentCloud and reviewed by The Verge. Rates are percentages of qualifying linked-out purchase value. The table does not estimate payment-processing expense, taxes, refunds, or net revenue because those inputs vary by developer. The proposals are not final court-approved fees.
This is more than a pricing detail. Program definitions become inputs to unit economics. Finance teams will need to model threshold risk, product eligibility, renewal treatment, refunds, cross-device purchases, and attribution duration. Product teams will need to decide which users see external offers and whether lower prices or added value are sufficient to overcome checkout friction.
The legal opening may still increase competition, but through margin, data, and product design.
It would be wrong to conclude that a nonzero fee makes external checkout meaningless. A developer saving ten or fifteen percentage points can fund a lower price, better bundle, customer service, or more acquisition. Owning billing data can improve dunning, recovery, segmentation, and support. Alternative payment methods can reduce cost or reach customers who do not want to use Apple’s billing.
The contrarian point is narrower: competition will not be measured by the existence of a link. It will be measured by the effective take rate and by the friction between app discovery and web conversion.
The court’s earlier findings demonstrate why. Apple’s former 27% rate left too little room after external payment costs, while link-design restrictions reduced conversion. A formally available path can remain commercially irrelevant when price and interface work together to suppress adoption. Conversely, a moderate fee with low-friction linking and clear customer value could create real competition even if it falls short of fee-free distribution.
Implications for operators
Model four separate economics: App Store billing, linked web checkout, web acquisition without an app touchpoint, and alternative app-store distribution where available. Do not compare commission rates in isolation. Include processing, fraud, tax, refunds, customer support, engineering, conversion loss, dunning performance, and the value of first-party billing data.
Treat attribution as a contractual and systems problem. The decisive questions include how long Apple’s claim lasts, which device or account events connect a web purchase to an app interaction, how cancellations and refunds reduce the fee base, and what audit data developers must provide. A rate that appears manageable can become expensive if attribution is broad or reconciliation is costly.
Design the offer around a reason to move. Users will not leave a familiar in-app flow simply because a developer prefers web checkout. Price differences, bundles, account portability, gifting, family administration, or payment choice must compensate for the extra step. Measure completion, not link clicks.
Preserve optionality in the billing stack. Product catalogs, entitlement systems, identity, taxes, and subscription state should not assume a single processor. The value of the legal opening rises when a developer can route customers without rebuilding the service layer.
Finally, avoid promising savings before the rules settle. Apple’s proposal is the start of another evidentiary phase, not a new tariff schedule. Epic has an opportunity to oppose it, and the court will determine what theory and rate, if any, are permissible.
Risks & open questions
The central risk to this thesis is legal. The court could limit Apple to demonstrable necessary costs and reject percentage commissions tied to platform value. Apple’s own characterization of the narrow cost base as essentially zero gives Epic a powerful argument that the proposed rates exceed the remand’s limits.
The thesis would also weaken if developers adopt linked checkout at high rates under a low or zero fee and Apple does not replace lost commission through other charges, eligibility rules, or interface friction. That outcome would show that payment separation can genuinely break the platform’s revenue claim.
Several operating questions remain unresolved: the attribution window; treatment of purchases made on another device; renewal duration; refund netting; audit and reporting requirements; the boundary between a link and general brand marketing; and whether program rates can change unilaterally.
There is also a customer-trust question. External checkout can improve price and choice, but it fragments billing support and refund expectations. Developers that gain control of the transaction also inherit the obligation to make that control useful and understandable.
Appendix / methodology notes
This report is based on Apple’s August 13, 2026 remand proffer, the Ninth Circuit’s December 11, 2025 opinion and subsequent orders, and contemporaneous reporting that reviewed the filing. Apple’s filing states its preferred legal and economic position; it is not a neutral cost study. The Ninth Circuit opinion is controlling procedural context but does not establish the final fee.
The chart-ready table reproduces proposed percentage rates and categorical treatment. It does not infer transaction volumes or developer savings. Any savings example would require developer-specific inputs: current App Store commission, gross linked-out purchase value, payment-processing rates, tax and fraud costs, refund rates, web-checkout conversion, and incremental operating expense.
The thesis is falsifiable. Evidence against it would include a final ruling that permits only a fixed cost-recovery charge, sustained high adoption of external checkout with no new platform toll, and materially higher developer-retained lifetime value after accounting for conversion and operating costs.