Apple v. Epic at SCOTUS (2026): The Case That Cannot Set Your Link-Out Rate

What Was Actually Filed, and What It Says

Between 16 and 21 September 2026 the Supreme Court’s public docket in No. 25-1311 recorded eight amicus briefs: the Washington Legal Foundation (16 Sep), the International Center for Law and Economics (19 Sep), and then six on 21 September — the Computer & Communications Industry Association and others, Professors Morley and Hessick, Protect the First Foundation, the Chamber of Progress, the Chamber of Commerce of the United States, Former Federal Antitrust Enforcers, and the United States.

The last one is the one worth reading, and its cover page is the first correction to the week’s coverage. It is titled “Brief for the United States as Amicus Curiae Supporting Neither Party.” Signed by Solicitor General D. John Sauer, it makes a three-part argument and asks for a three-part result:

• Contempt improper on the 27% commission. “A fair ground of doubt exists whether the injunction’s text, read in context, barred the 27% commission: the anti-steering ban neither mentions nor covers commissions.”
• Contempt proper on the plain buttons. “The ‘plain buttons’-only restriction clearly violates the injunction’s express terms.” The government does not ask for this to be disturbed.
• Remand on everything else — the four other design restrictions, and the bad-faith findings, on which “the government takes no position.”

The conclusion is one sentence: “The judgment of the court of appeals should be reversed in part, affirmed in part, and vacated in part.” That is not the same document as the one described in headlines about briefs showing strong support for Apple’s arguments. On the government’s own view, the part of the case that governs how your link or button may look survives; the part that governs what Apple may charge goes back down.

The Question Presented Is Not About Money

On 30 June 2026 the Court granted certiorari “limited to Question 1 presented by the petition.” Question 1 asks whether a court may hold a party in civil contempt for violating an injunction’s “spirit” where the injunction is silent as to the conduct at issue, or whether contempt must rest on an order that clearly and unambiguously proscribes the precise conduct.

That is a question about the law of civil contempt. It is a genuinely important one — it is why the United States has an interest at all, since the federal government is routinely both bound by and the beneficiary of injunctions — but it is not a question about App Store economics. The remedies available to the Court are to affirm, reverse or vacate contempt holdings and send them back. There is no version of a ruling in which the Court announces a percentage.

The Ninth Circuit made the same point in plainer language on 28 April 2026, when it refused to keep its mandate on hold: “Even if the Supreme Court agrees with Apple’s arguments, there would still be further proceedings on remand, particularly on the question of commission, and those proceedings are likely to look similar, if not the same, regardless of certiorari.” Apple quotes that sentence in its own filings in order to call it “clearly incorrect.” Whether or not Apple is right about that as a matter of litigation strategy, it is the disagreement that tells you where the rate actually gets decided.

Two Proceedings, Only One of Which Sets Your Rate

Almost every confusion about this story dissolves once the two tracks are separated:

So the honest headline for 21 September is narrow: the United States filed a brief on a contempt-standard question in a case that will not be argued this calendar year. The number that touches your P&L is being litigated a floor down, on a schedule that has already produced orders in August.

The August Fight Nobody Covered This Week

This is the part of the record most worth knowing, and it is absent from the amicus-brief coverage because it happened five weeks earlier.

After certiorari was granted, Apple asked the district court to pause the remand until the Supreme Court resolved the contempt question. On 11 August 2026 the district court refused (Dkt. 1706) and, in Apple’s own account, ordered it to submit an evidentiary proffer by 6:20 p.m. ET the next day justifying a commission on link-out purchases “under the very Ninth Circuit decision this Court granted certiorari to review.”

Apple went straight up. On 12 August 2026 it filed an application (No. 26A194) directed to Justice Kagan to stay the lower-court proceedings. An administrative stay was entered the same day. Epic opposed that evening. Apple replied the next morning. On 13 August 2026 the application was denied.

That is the single most decision-relevant fact in the whole docket for anyone running a US app business: Apple asked the Supreme Court to freeze the rate-setting proceeding, and the Supreme Court said no. The proceeding that can change your commission is live, and it has been live all through the amicus season that made this week’s headlines.

Apple’s proffer, filed on 13 August and reported at the time, proposed a tiered structure: 15% for standard apps, 10% for the Video, News and Mini Apps Partner Programs and for subscription renewals, and 5% for Small Business Program apps. Those are a party’s proposal, not an order. They are useful only as the top of the plausible range for planning — the number that binds you is whatever Judge Gonzalez Rogers signs.

The Dated Record

Every row below is from the public Supreme Court docket for No. 25-1311 or from the filings on it:

The 3% That Both Sides Are Really Arguing About

Apple got to 27% by taking its 30% in-app rate and applying what it called a “cost of payments discount” of 3% — the reasoning being that it no longer processes the payment, but still supplied the platform and the user who made the purchase. The district court’s finding, which the Ninth Circuit repeated, was that Apple “knew that developers would almost always incur additional costs for off-app purchases exceeding 3%.”

That is not an abstract legal point. It is the same line every finance team draws when it models a web checkout: card processing, fraud, chargebacks, refunds, dunning, sales tax and VAT determination, and the engineering to run all of it. If that stack costs more than 3%, then 27% is worse than 30% and the link-out is theatre.

The district court had a number for how that played out in practice. Of approximately 136,000 developers on the App Store, only 34 applied to use external purchase links in the four months after the policy rolled out. Whatever one thinks of the contempt analysis, that adoption figure is the cleanest available evidence of what a commission set near parity does to link-out behaviour — and it is the reason the eventual number matters far more than which way a contempt holding goes.

What a Rate Change Does to Your Subscription Stack

Here is the part that is specific to the tools rather than the case, and it runs against intuition.

The mobile subscription platforms in our subscription analytics directory mostly bill as a percentage of revenue — but the base is gross, before Apple’s cut. From our verified listings: RevenueCat’s Pro plan is 1% of monthly tracked revenue “pre-platform-cut” past $2,500 MTR; Adapty is 1% of monthly revenue “measured in USD before platform cut” past $5,000. So the same gross sale produces the same fee whether Apple takes 30%, 15% or nothing at all. What moves is the fee’s weight against what you actually bank:

Three consequences follow, and none of them is in the litigation coverage.

First, a rate rise does not reduce your tooling bill. If the district court lands on 15%, your analytics line stays at $1,000 while your receipts fall by $15,000. Anyone who built a model assuming the percentage-of-revenue vendor absorbs part of a commission change has the mechanics backwards.

Second, the vendors are not interchangeable on this axis. Superwall bills 1% only on paywall-attributed revenue and gives the subscription infrastructure away at any scale, so its fee tracks a narrower base by construction. Qonversion starts at 0.6% on its Starter plan and 0.8% on Growth. We worked the four denominators against each other in RevenueCat vs Adapty vs Superwall vs Qonversion: The Denominator, and the ranking genuinely changes with your revenue mix — which is exactly what a commission order would change.

Third, the question to ask your vendor now is a boundary question, not a price question. If you run both IAP and a linked-out web checkout, ask in writing which dollars count as “tracked revenue” on your plan, and whether web-billing revenue is metered on the same basis as store revenue. That answer is contract-specific, it is the thing that determines your bill under any new rate, and it is cheap to get today and expensive to discover later. If your free tier is doing the work, the exact ceiling on each free plan is where to check whether a mix change pushes you over it.

What to Do Before the Rate Lands

Three cases cover most readers.

1. You already run US link-outs. Nothing changed this week; the rate is still 0%. The thing worth doing is building the 15% case into your model now rather than after an order, because Apple’s own proffer is the best available ceiling and it is public. Run your linked-out cohort at 0% and at 15% and see whether the web checkout still beats IAP once your processing, tax and support costs are in — that is the same 3% argument the courts have been having, applied to your numbers. Our profit margin calculator is enough for the first pass.

2. You have been waiting for the dust to settle before building link-out at all. The dust will not settle on the Supreme Court’s schedule — Epic’s brief is not due until 13 November 2026 and there is no argument date. The district court will almost certainly move first. If the build is otherwise justified, the litigation is not a reason to wait; the commission uncertainty is a reason to build the accounting so the rate is a parameter rather than an assumption baked into a hundred places.

3. You are choosing a subscription platform this quarter. Do not let agentic-sounding roadmaps or link-out positioning drive the pick. Score the denominator, the free-tier ceiling, and the contractual definition of tracked revenue — the three things that will still be true whatever the order says. Vendor stability belongs on that list too, for reasons we set out in SaaS Vendor Shutdown Risk. And if you are integrating Apple sign-in alongside billing, the three-domain change is a separate, live piece of Apple housekeeping worth closing off.

What This Piece Does Not Claim

This is not a prediction. Nothing here says how the Supreme Court will rule, what rate the district court will set, or when. The eight amicus briefs filed between 16 and 21 September are not evidence about an outcome; amicus support is not a vote count, and the Court granted review on a contempt-standard question that it could resolve in several directions without saying anything about App Store commissions.

Nor is it legal advice. Every date, quotation and case number above comes from the public Supreme Court docket in No. 25-1311 and from documents posted on it — principally the Brief for the United States as Amicus Curiae Supporting Neither Party (21 September 2026) and Apple’s Application to Stay Lower-Court Proceedings (12 August 2026). Where a fact comes from reporting rather than a document on that docket, it is flagged: the 15% / 10% / 5% tier figures are from contemporaneous coverage of Apple’s 13 August district court proffer, which is not posted on the Supreme Court docket, and should be treated as reported rather than verified here.

Finally, this changes no listing. RevenueCat, Adapty, Superwall and Qonversion are scored in our directory on published pricing, billing basis and free-tier ceilings, none of which moved. If a commission order changes how any of them meters revenue, the listings get updated then, against the vendors’ own pricing pages rather than against a court filing.

FAQ

What does Apple charge on US link-out purchases right now?

Zero. The Ninth Circuit provided that the rate “would remain at zero until the district court sets a new rate,” and nothing filed in September changed that. The number changes by order of the Northern District of California in case 4:20-cv-05640.

Did the US government side with Apple?

The brief is captioned “supporting neither party” and asks for the judgment to be reversed in part, affirmed in part and vacated in part. It would undo contempt on the 27% commission and leave contempt standing on the “plain buttons” restriction.

Can the Supreme Court set the commission rate?

No. Certiorari was granted limited to Question 1, a civil-contempt standard question. Its remedies are to affirm, reverse or vacate and remand.

When is this decided?

Epic’s merits brief is due 13 November 2026 and no argument date appears on the docket. The district court track is the faster one and is unstayed — Justice Kagan denied Apple’s stay application on 13 August 2026.

Will a higher Apple commission lower my analytics bill?

No. RevenueCat and Adapty both meter on gross revenue before the platform cut, so the fee is unchanged by Apple’s rate while your receipts fall. On $100,000 gross, a $1,000 fee is 1.00% of receipts at 0% and 1.43% at 30%.

Sources: Supreme Court of the United States, docket for Apple Inc. v. Epic Games, Inc., No. 25-1311 (all filing dates, the 30 June 2026 grant “limited to Question 1,” the 23 July 2026 briefing order setting 14 September and 13 November 2026, and the 11–13 August 2026 entries for Application 26A194). Brief for the United States as Amicus Curiae Supporting Neither Party, No. 25-1311, filed 21 September 2026 (caption; the “fair ground of doubt” argument; “the anti-steering ban neither mentions nor covers commissions”; “the ‘plain buttons’-only restriction clearly violates the injunction’s express terms”; the conclusion; the 30% IAP rate, the 3% “cost of payments discount,” and the district court’s finding that 34 of approximately 136,000 developers applied to use external links in four months). Application Directed to the Honorable Elena Kagan to Stay Lower-Court Proceedings, No. 26A194 (25-1311), filed 12 August 2026 (the 11 August 2026 district court order, Dkt. 1706; the 6:20 p.m. ET proffer deadline; the “between 12% and 27%” link-out rates; the Ninth Circuit’s 28 April 2026 stay denial, 2026 WL 1215632; and the quoted holding that Apple’s rate “would remain at zero until the district court sets a new rate”). Case history: 559 F. Supp. 3d 898 (N.D. Cal. 2021); 67 F.4th 946 (9th Cir. 2023); 144 S. Ct. 681 and 682 (2024); 161 F.4th 1162 (9th Cir. 2025). The 15% / 10% / 5% proposed tiers are from 9to5Mac’s 13 August 2026 report on Apple’s district court proffer and are identified above as reported. Pricing and billing-basis figures for RevenueCat, Adapty, Superwall and Qonversion are from ToolMasta’s own verified directory listings. No prediction is made about the outcome of either proceeding, and nothing here is legal advice.