

Epic v Apple reaches the Supreme Court, and web checkout waits on the clock
On 14 September 2026 Apple filed its opening brief in the US Supreme Court, asking the justices to vacate a contempt finding that has kept its commission on US link-out purchases at zero since 30 April 2025.
By Katie Delaney / 2026-09-20 / 16 min read

What Epic v Apple asks the Supreme Court to decide
Nobody outfoxes a courtroom calendar, and the fox knows it. The patient prowl this week leads not to a ruling but to a schedule, because Epic v Apple is now a Supreme Court case with a docket number (25-1311), one live question and, when folkfox read the docket on 20 September 2026, no argument date. Apple filed its opening merits brief and the joint appendix on 14 September. For an app marketer the useful question is not who wins. It is how long the zero-commission trail through the thicket stays open.
The Epic v Apple Supreme Court docket records a petition for certiorari filed on 21 May 2026 and granted on 30 June, limited to Question 1. The Court's question-presented sheet words that question plainly: may a court hold a party in civil contempt for violating an injunction's "spirit" where the injunction is silent on the conduct at issue, or must contempt rest on an order that clearly and unambiguously proscribes the precise conduct? Apple's second question, about Trump v. CASA and an alleged antitrust exception, was not taken. The apple app store lawsuit, filed in the Northern District of California under case number 4:20-cv-5640, has been narrowed to a single procedural quarry.
The same docket shows the Court extending the schedule on 23 July: Apple's brief and appendix by 14 September, Epic's brief by 13 November. Supporting briefs are arriving already, with the Washington Legal Foundation filing on 16 September and the International Center for Law and Economics submitting on 19 September. There was a sidelight in August too. Apple applied on 12 August (application 26A194) to stay the district court's remand proceedings; Justice Kagan entered an administrative stay until 5 p.m. Eastern on 13 August, then denied the application the same day.
| Item | Value |
|---|---|
| 2025-04-30 | No-fee order |
| 2025-12-11 | Appeal ruling |
| 2026-06-30 | Cert granted |
| 2026-08-13 | Apple offers 15% |
| 2026-09-14 | Apple's brief filed |
| 2026-11-13 | Epic reply due |
Read that line the way a fox reads a hedgerow: the gaps say more than the marks. Two dates that circulate in coverage need care. 9to5Mac and TechTimes both put Apple's reply on 14 December, and TechTimes adds amicus windows in September and November, but the Court's own docket lists none of those dates, so folkfox treats them as reported rather than confirmed. The 9to5Mac report on Apple's brief also says the case could be argued as early as January 2027; TechTimes calls that the earliest plausible slot and expects a final ruling no sooner than June 2027. Those are forecasts from trade press, not calendar entries.

What is Apple actually arguing? In the words 9to5Mac reports from the brief, civil contempt requires "clear notice that the conduct at issue is proscribed". The Mac Observer's summary adds that the injunction behind the contempt finding ran to 75 words and never mentioned commissions. In Epic v Apple the justices are not being asked to name a fee, and that is the detail app teams tend to miss. Naming the number is a job for the district judge in Oakland, working on a separate track.
Where the Apple external payment fee stands after the Ninth Circuit
The Ninth Circuit opinion of 11 December 2025 is the document to keep on the desk, because it is the apple epic ruling that changes budgets. The panel held, in its opinion, that Apple's 27% link-out commission "had a prohibitive effect". It also refused to bless the district court's blanket ban: the zero-commission rule, the panel said, was "more like a punitive criminal contempt sanction" than a civil one. So the appeals court kept the contempt finding and sent the fee back to be set again.
The court then sketched what a lawful fee might look like. Apple should be able to charge a commission on linked-out purchases based on costs that are "genuinely and reasonably necessary" for coordinating external links, "but no more", and it is entitled to some compensation for the intellectual property used to complete those purchases. The opinion also lays bare how the 27% figure was built: Apple took its 30% in-app commission and applied a "cost of payments discount" of 3%, applied for seven days after the tap. The court recorded that Apple believed the discount was too small to produce meaningful "developer adoption of link-out". A fee designed to be ignored was, unsurprisingly, ignored.
Now the remand in Epic v Apple. According to 9to5Mac's reading of the 13 August filing, Apple proposed 15% on link-out purchases for standard apps, 10% for the Video, News and Mini Apps Partner Programs and for subscription renewals, and 5% for Small Business Program members. Apple argued the rates would let developers "link out profitably" while recovering some compensation for its intellectual property. Epic replied that under the Ninth Circuit's definition of necessary costs it would charge 0%, a position TechCrunch also reported. None of this is a ruling. It is opening positions in the Epic v Apple fee fight.
One more scent worth following. Apple also asked the district court to refer the dispute to a settlement conference, and Epic's filing of 17 August said it remains "willing to entertain any direct approach from Apple" with a serious proposal while doubting a referral would help. Settlement is possible, and it would rewrite the fee without a judge or a justice. It is also the outcome nobody can plan around, which is the point of the next section.
What does Apple's own rulebook say today, while Epic v Apple runs? The App Review Guidelines are blunt on the United States storefront and restrictive everywhere else, and the fee itself appears nowhere in them.
| Where it applies | What the guidelines say | Rule |
|---|---|---|
| United States storefront | Entitlements are not required for buttons, external links or other calls to action | 3.1.1(a) |
| Other storefronts | Buttons and links are barred unless a regional StoreKit External Purchase Link Entitlement applies | 3.1.1(a) |
| Music streaming apps | Regional Music Streaming Services Entitlements allow a link, including a buy button | 3.1.1(a) |
| Reader apps | External Link Account Entitlement allows an informational account link | 3.1.3(a) |
| Link-out fee | No figure appears in the published text; the number lives in the court record | None |
- United States storefrontEntitlements are not required for buttons, external links or other calls to action3.1.1(a)
- Other storefrontsButtons and links are barred unless a regional StoreKit External Purchase Link Entitlement applies3.1.1(a)
- Music streaming appsRegional Music Streaming Services Entitlements allow a link, including a buy button3.1.1(a)
- Reader appsExternal Link Account Entitlement allows an informational account link3.1.3(a)
- Link-out feeNo figure appears in the published text; the number lives in the court recordNone

The margin arithmetic: web checkout at 0%, 15% and 27%
Epic v Apple fees are arithmetic before they are strategy, so start with the arithmetic. Take a $10 monthly subscription, Apple's standard 30% in-app commission, and a web checkout that costs 2.9% plus 30 cents in card processing, the Stripe rate RevenueCat's own analysis uses. This is illustrative and deliberately simple: it ignores tax, refunds, chargebacks and any billing-software fee. In-app, the developer keeps $7.00. On the web with no Apple fee, it keeps $9.41. With Apple's proposed 15% it keeps $7.91. With the old 27% it keeps $6.71, which is less than the in-app route.
| Item | Value |
|---|---|
| Web at 0% fee | 7 to 9.4 |
| Web at 15% fee | 7 to 7.9 |
| Web at 27% fee | 7 to 6.7 |
| Small firm, web at 5% | 8.5 to 8.9 |
The chart hides a sharper number. Divide in-app net by web net and you get the share of in-app conversion the web funnel must keep to break even: about 74% at no fee, 89% at a 15% fee, and above 100% at 27%. A web route can lose up to a quarter of its buyers and still break even today. The same route can lose no more than about one in nine if Apple's 15% proposal is adopted, and it cannot win on price alone at 27%. Fee risk, in other words, is conversion risk in a costume.
What have vendors measured? Superwall's early data, published on 7 May 2025, showed initial conversions falling about 11.9% while net proceeds rose roughly 19.85%, in Superwall's app-to-web study, which itself warned that refund and chargeback rates needed more time to mature.
Adapty's one-year review of app-to-web paywalls, published on 25 June 2026, reports one app whose purchase conversion fell from 4.95% to 3.69% (down 26%) while revenue per paywall viewer rose 53%, because the team built a different web offer with longer plans rather than copying the native weekly price. A Small Business Program app in the same review saw purchase conversion rise 55% but revenue per viewer slip 9%, and the team walked the test back. Adapty calls its read directional rather than a controlled 50/50 test.
RevenueCat, which offers web checkout tooling itself and still argues most apps should hesitate, published its own experiment in an April 2026 post. In RevenueCat's write-up, conversion to paying moved from 7.8% to 7.4% (down 5.1%), subscribers set to renew rose from 46 to 101, and refunded customers rose from 28 to 38. Those are small counts from a single app, so the refund jump is a warning light rather than a rate. Treat every vendor number here as a vendor number, from one app or a handful, and dated.
| Item | Value |
|---|---|
| Adapty app 1, 2026 | 74 of 88.5 |
| Superwall test, 2025 | 88.1 of 88.5 |
| RevenueCat test, 2026 | 94.9 of 88.5 |
A fee you can model is an annoyance. A fee you never modelled is a hole in the budget.
Notice what the fee arithmetic cannot see. Adapty's paywall test won because the web offer was redesigned, not because the rail changed, which means the 53% gain is a product result as much as a fee result. Run the yardstick against your own funnel and refund history before Apple's figure moves.
Web to app funnels versus link-outs: which one the case touches
Two different journeys hide under one label, and Epic v Apple touches only one. Adapty separates them cleanly: app-to-web is a button inside an installed app that opens a browser checkout, which is what the April 2025 order in Epic v Apple enabled for United States iOS apps, while web to app starts on the web, usually from an ad, runs through a quiz and paywall, takes payment and only then sends the buyer to the install. Web to app, Adapty notes, "has been around for years, predates the Epic ruling, and works on any platform in any geography". The Adapty guide keeps its data on the first journey.
That distinction matters for planning. A web to app funnel earns its keep on acquisition economics and does not depend on Apple's link-out fee; a court-set commission changes the return leg, the upsell, the renewal and the win-back that a loyal user reaches by tapping a link inside the app. The prize for the first journey is cheaper acquisition. The prize for the second is a bigger margin on people you already own. The fox would fund both, but only one of them carries a legal clock.
The rulebook is still enforced while the lawyers argue. Adapty's account is that Apple briefly removed the calorie-counting app Cal AI in April 2026 over web payment violations, including offering Stripe as the only payment option and burying renewal terms behind a trial toggle; treat that as Adapty's reading of the episode, not a statement from Apple. The practical lesson is the one the guidelines already imply: keep in-app purchase available, describe the redirect plainly, and expect App Review to keep looking.
Apple, meanwhile, is not conceding the checkout. Its developer news of 16 September 2026 announced Bundles and Suites for iOS 27, multiseat purchases enabled by default for subscriptions in App Store Connect, Volume Purchasing launching on 22 October 2026 and Group Purchases arriving this winter. Apple is making the in-app till more attractive at the same moment it defends the fee, and a subscription team that has one eye on web margin should keep the other on those new in-app selling surfaces.
Who feels this most? The reader is an app marketer, often in a regulated subscription business, and each vertical carries a different snag. Fintech apps inherit a financial-promotion problem, because the web checkout becomes a regulated surface that compliance must approve, as the Nubank rate-first launch showed for claims. Health apps should think carefully about what data a browser form collects. iGaming-adjacent apps must rebuild age gates and spend limits on the web page rather than assume Apple's flow carries them. Music apps already have entitlement routes in the guidelines. Cybersecurity brands sell on trust, and Adapty warns operators to watch reviews for words like scam or deceptive when a checkout looks unfamiliar.
Attribution deserves its own moonlit hour. When the purchase moves off StoreKit, the sale is recorded on your server, so your measurement partner has to receive it from you. The measurement partner costs of Apple ATT and the ATT alternative prompt brief explain why signal was already thin before a web checkout split the funnel in two.

A plan that survives both outcomes: durable, not permanent
Here is the folkfox position, stated without a hedge. Build web-to-app funnels and web checkout after Epic v Apple as durable through at least early 2027, but not as permanent. The Court has no argument date on its docket, the briefs run into December, and TechTimes expects no final ruling before June 2027. Durable, then. Not permanent, because the Oakland track could set a fee before Washington rules, Apple has already put 15%, 10% and 5% on the table, and a win for Apple in the Supreme Court could reopen the sanctions the contempt finding supported. That last point is folkfox's reading, not a holding.
So plan the economics for both outcomes. Model your subscription at zero commission, at a cost-based fee in the 5% to 15% band Apple has proposed, and at a restored 27%. Keep the in-app purchase paywall as the default and treat web checkout as an extra placement, never a replacement. In Epic v Apple terms, that fallback is cheap insurance, it matches the guidelines, and it means a fee change becomes a spreadsheet update rather than a rebuild. Then choose the one number to defend: proceeds per paywall viewer, not conversion rate, which is also Adapty's advice.
Measure three things on the web side from day one. Conversion, first, split by placement, because Adapty's companion app saw a nearly fourfold spread between its best and worst placements. Refunds, second, because vendor data already flags them as the cost that arrives late. Attribution, third, because a sale that never touches StoreKit is invisible to your usual reports until you pipe it through.
If you are still shaping the wider programme, our app marketing service covers the funnel and the measurement, and the App Store submissions checklist and app store optimisation launch checks cover the store-side work that still decides who reaches the paywall at all. If you want a second pair of eyes on the numbers, talk to folkfox and bring your own conversion and refund data.
Frequently asked questions
What is Epic v Apple about at the Supreme Court?
Epic v Apple, docket 25-1311, asks one narrow question: whether a court can hold a party in civil contempt for violating an injunction's spirit when the order is silent on the conduct. Apple filed its opening brief on 14 September 2026. The Court has not been asked to set a commission.
Is there an Apple external payment fee on US link-outs right now?
TechTimes reported on 15 September 2026 that US developers have operated under a zero-commission baseline since 30 April 2025. No final fee exists. Apple proposed 15% for standard apps, 10% for partner programmes and renewals, and 5% for the Small Business Program on 13 August, and Epic argued for 0%.
What did the Apple Epic ruling from the Ninth Circuit change?
On 11 December 2025 the Ninth Circuit upheld the contempt finding but said the blanket zero-commission rule read like a punitive sanction. It allowed Apple to seek a fee based on costs genuinely and reasonably necessary for coordinating external links, and sent the number back to the district court.
Does the Apple App Store lawsuit affect web to app funnels?
Mostly not directly. Adapty describes web to app funnels, which start on the web and end with the install, as predating the ruling and working in any geography. The case governs the in-app link-out route, which is what a court-set fee would reprice.
When will the Supreme Court decide the Apple and Epic case?
Not soon. Epic's response is due on 13 November 2026 and the docket lists no argument date. TechTimes expects argument no earlier than January 2027 and a final ruling no sooner than June 2027. Those are press forecasts, not court entries.
Should a regulated subscription app use web checkout?
Test it as an added placement, not a replacement. Keep in-app purchase as the default, get compliance to approve the web page, and track proceeds per paywall viewer, refunds and attribution. Model the economics at zero commission and at a restored fee before you scale.
Mobile Measurement Partner: 3 Hidden Costs of Apple ATT
Why attribution signal was already thin before web checkout split the funnel in two.
Read the pieceAPP marketingApple ATT alternative prompt: an essential EU app brief
How EU app teams test permission journeys without inventing uplift.
Read the pieceAPP marketingApp Store submissions: 5 practical checks
Five checks before you submit for the latest Apple OS releases.
Read the pieceFintechFintech Digital Marketing Agency: 5 Sharp Nubank Lessons
A claims-risk map of Nubank's US launch, useful for any regulated subscription page.
Read the piecePlan the web checkout before the fee is decided
folkfox builds app funnels, paywall tests and measurement that hold up at zero commission and at a restored fee. Bring your conversion and refund data.
Want folkfox in your Google results and AI answers? Set folkfox as a preferred source.

