Android Distribution Went Two Ways in 2026: Rival Stores Opened, Sideloading Closed Up

Android distribution changed twice in 2026: rival app stores now sit inside Google Play in the US, while unverified sideloading gained a 24-hour install delay. Both changes land on anyone licensing or publishing APK games, and they pull in opposite directions.

If you distribute Android games and you only followed one of this year's two big stories, you have half the picture — and probably the wrong half. The story that got the headlines was the Play Store opening up. The story that will actually reshape how APK catalogues move around the world is the one that got far less attention: Google is closing the unverified install path, market by market, starting this September.

One of those changes widens your options. The other narrows them. They are not going to cancel out, and the practical consequence for a licensee is that where your APK can legally and technically land is now a question you have to re-answer per territory, not once.

🚪 On 22 July, the Play Catalogue Stopped Being Play's Alone

On 22 July 2026, Google switched on two programmes at once in the United States: the Play Catalog Access Program and the Third-party App Store on Play Program. Per Google's Play Console Help documentation, developer app and game listings distributed in the US — names, icons, descriptions, screenshots and videos — are made available to enrolled third-party US Android app stores, which can then offer those apps to their own users.

Two details matter more than the headline.

First, enrolment was the default. Google notified developers on 22 June 2026 that unless they opted out by 22 July, their listings would be included. Anyone who did not open Play Console and make an explicit choice in that window is in the programme now. The control lives under Settings > Catalog Settings, with three options: publish all listings on all third-party stores, manage each store individually, or publish none. That is adjustable at any time, but the default position is worth knowing if you have not looked.

Second, Google still runs the checkout. The download completes through Google Play on the same terms as any other Play download, and Play's service fee continues to apply. A rival store becomes a discovery surface — a shop window onto the same warehouse. It is not an independent distribution channel with independent economics. If you were modelling third-party stores as a way around the service fee, that model does not survive contact with the documentation.

How it got here

This is court-ordered, not voluntary. It descends from the 2024 injunction in Epic Games v. Google. Google settled with Epic in November 2025 and proposed a modified remedy in March 2026 — a "Registered App Stores" scheme where rival stores would install through a registration flow rather than from inside Play. On 15 July 2026, as MacRumors and TechTimes both reported, Google and Epic jointly withdrew that proposal after a court-appointed economic expert questioned its pro-competitive effect, leaving the original injunction in force. A week later the programmes went live as originally ordered.

Read that sequence as a signal about durability. This is not a product decision Google can quietly reverse next quarter.

🔒 The Other Direction: Unverified Installs Get a 24-Hour Lock

Now the change pulling the other way. Google announced in August 2025 — TechCrunch broke it on 25 August — that developer verification would be required for apps installed on certified Android devices outside the Play Store. Sideloading does not disappear. It gets a gate.

The timeline, per Android Authority's reporting on Google's rollout schedule:

• June 2026 — the system service that enforces this begins rolling out to devices.
• August 2026 — limited-distribution developer accounts launch, allowing an app to be shared with up to 20 devices without a government-issued ID or developer fee.
• 30 September 2026 — requirements activate in the first markets: Brazil, Indonesia, Singapore and Thailand.
• 2027 — expansion across certified Android devices globally.

When an app comes from an unregistered developer, the install no longer happens with a tap. The advanced flow adds multiple steps and a mandatory 24-hour lock before installation completes. That is deliberate friction, and it is aimed at malware — but friction does not distinguish between a malicious APK and your perfectly legitimate branded game handed to a client on a USB stick at a kiosk build.

Look hard at that first-market list. Brazil, Indonesia, Singapore, Thailand. Those are not incidental territories for APK licensing — they are among the markets where alternative distribution, operator app stores and OEM channels carry the most weight. Google is starting enforcement precisely where the non-Play install path is busiest.

The verification programme is not Google-only either. Android Authority lists the participating stores as Google Play, Samsung Galaxy Store, Xiaomi GetApps, HONOR App Market, OPPO App Market, vivo V-Appstore and Palm Store. If your distribution plan leans on OEM stores as the Play alternative, those stores are inside the same verification perimeter.

💵 The Fee Card Was Rewritten in the Same Window

Buried under both stories, Google separated its service fee from its billing fee. From 30 June 2026 in the EEA, UK and US, per Google's service fee documentation:

• The billing fee is 5% in the US, UK and EEA. It applies to Google Play Billing and to alternative billing — but not to external web links.
• The first $1M in annual earnings is charged at 10% + billing fee across all transaction types.
• For new installs at standard rates: 20% + billing fee on non-recurring transactions, 10% + billing fee on recurring.
• For existing installs: 25% + billing fee on non-recurring in-app, 20% on a non-recurring external web link, 10% + billing fee on recurring.
• Developers in the Apps Experience Program or Games Level Up Program pay 15% + billing fee on new-install non-recurring transactions, and 20% + billing fee on existing-install in-app.

The rollout is staged: 30 June 2026 for the EEA, UK and US; 30 September 2026 for Australia and Japan; 31 December 2026 for Korea; 30 September 2027 for the rest of the world.

The line worth internalising is the gap between an in-app purchase on an existing install and the same purchase through an external web link — 25% + 5% against 20% with no billing fee. That is a ten-point spread, and it is the first time Android has offered a real reason to build a web checkout. Whether ten points is worth the conversion loss of pushing a player out to a browser is an empirical question for your title, not a foregone conclusion. Most teams overestimate the gain and underestimate the drop-off.

One more date: developers in the external content links and alternative billing programmes must begin reporting transactions and paying applicable service fees from 1 October 2026. If you enrolled in the autumn of 2025 and have been enjoying an unbilled window, that window closes in seven weeks.

🌍 If You Are Not Selling Into the US, Only Half of This Applies

Worth saying plainly, because it gets muddled: the catalogue-sharing programmes are US-only. They came out of a US court order. A publisher operating in MENA, SEA or LATAM gets no new discovery surface from them.

But verification is global on a delay, and its first four markets are outside the US entirely. So the asymmetry for a non-US operator is unhelpful: none of the opening-up, all of the locking-down, and the locking-down arrives in your territory first. If your Android strategy depends on direct APK delivery — an operator portal handing out installs, an OEM preload pipeline, an enterprise fleet, a low-connectivity market where a downloadable file beats a store listing — September is your date, not 2027.

🧾 What This Changes in a Licence Agreement

Here is the part most buyers will miss. Distribution changes of this size make licence scope the expensive clause, and a lot of APK licences were written before any of this existed.

Things to check in a contract you are about to sign, or one you already signed:

• Does the licence name channels, or platforms? A licence granting "distribution on Google Play" is not obviously a licence to distribute through a third-party store that surfaces a Play listing. It probably is fine, since the download still completes through Play — but "probably fine" is not what you want in writing when a licensor decides otherwise later.
• Who is the publisher of record? Verification attaches to a developer identity. If your licensor publishes under their account and you resell, you have no verification standing. If you publish under yours, you are the verified party carrying the obligation. Both are workable. Neither is workable if nobody has decided which it is.
• Does it cover OEM and operator stores by name? Galaxy Store, GetApps, AppGallery, HONOR, OPPO, vivo and carrier stores are distinct channels. Vague "all Android distribution" language usually survives; enumerated lists written in 2023 usually do not.
• Territory language against a per-market rollout. Verification lands in four countries this September and elsewhere later. A licence with per-territory rights now interacts with a per-territory technical gate. Make sure the two lists are compatible.
• Source-code deals shift the burden entirely. If you hold source and build your own APK, you are the developer for verification purposes. That is more control and more compliance work — which is the whole trade in a source-code licence and always has been.

None of this is exotic. It is the same due diligence that has always separated buyers who read the grant clause from buyers who read the price. Our own notes on licence terms and pricing structures cover the general shape; the point here is that 2026 added a technical enforcement layer behind clauses that used to be purely commercial.

🚫 Four Ways Operators Will Get This Wrong

Treating rival stores as a new revenue channel. They are a shop window. The download and the fee still route through Play. Budget them as incremental discovery — a few percent of installs, maybe — not as a second P&L line. Anyone selling you a "Play alternative" that quietly still pays Play's service fee should be asked to show the maths.

Assuming opt-out was the safe default. It was not the default at all. If you did nothing in that June-to-July window, you are enrolled. That may well be what you want; decide it deliberately rather than discovering it in a quarterly review.

Building a 2027 direct-APK plan on 2025 assumptions. If any part of your distribution assumes a user can tap an APK and have it install, that assumption expires in Brazil, Indonesia, Singapore and Thailand on 30 September. Kiosk deployments, event builds, operator side-channels and enterprise fleets all need to be re-walked against the advanced flow.

Chasing the ten-point external-link spread without measuring conversion. Twenty percent with no billing fee beats twenty-five plus five only if the same number of players complete the purchase. They will not. Run it as an experiment on one title before rebuilding a checkout you cannot easily un-build.

🧭 What to Do Before 30 September

Concretely, in order:

1. Open Play Console, go to Settings > Catalog Settings, and see what position you are actually in. Five minutes.
2. List every route by which one of your APKs reaches a device without passing through a store. For each one, ask what happens when a 24-hour lock is inserted.
3. Identify who the verified developer is for every title you license — you or the licensor. Get it in writing if it is not already.
4. If you sell in Brazil, Indonesia, Singapore or Thailand, treat 30 September as a hard deadline rather than a milestone.
5. If your business runs primarily on HTML5 rather than APK, note that none of this touches you — which is itself a strategic data point worth weighing when you next choose a format for a new deployment.

That last point deserves a sentence more, because it is the quiet lesson of 2026. Every change described here is a change to who controls the install. HTML5 does not have an install, which is why browser-based catalogues keep absorbing use cases — telecom portals, in-flight systems, kiosks, signage — that Android used to own by default. That is not an argument for abandoning Android; Android is where the paying mobile audience is. It is an argument for knowing which of your deployments actually needed a native app and which inherited one out of habit.

🎮 Where Forestry Games Fits

We have licensed games since 2017 and the catalogue runs to 1,049 titles across HTML5 and Android APK, with titles also published on Google Play and the App Store. The practical value of a dual-format catalogue in a year like this one is optionality: when a distribution route changes under you, you can move a deployment to the format that still fits, rather than re-licensing from scratch. If you are re-checking your Android exposure ahead of September and want to talk through which of your channels are affected, that is a conversation worth having before the deadline rather than after.

📌 The Short Version

Rival app stores are real, they are US-only, and Google still takes the fee. Sideloading survives with a 24-hour penalty box, arriving in four major APK markets this September and everywhere by 2027. The fee card now separates service from billing and rewards external checkout by about ten points, if your conversion can absorb it.

The action item is not strategic, it is clerical: go and find out which of your distribution routes assume a frictionless install, and fix those first. Every operator who gets caught by this in October will have been caught by a route they forgot they had.