TL;DR
On July 15, the day before the evidentiary hearing that was meant to be the "final act," Epic and Google jointly withdrew their renewed motion to modify the Epic v. Google injunction. Google has confirmed it will keep complying with Judge Donato's original 2024 order. From July 22, rival app stores start appearing inside the US Google Play Store: distribution the original injunction always required and the settlement would have removed.
My read on why they folded: three days earlier, on July 12, the court's own independent economist, Prof Nancy Rose, filed a declaration that took the revised settlement apart. I think Epic and Google looked at it, did the math on their odds in front of an already-skeptical judge, and decided withdrawal beat a likely loss.
But this is not a clean win for developers, and I want to be precise about why. On the two big structural questions, store distribution and link-out rights, the withdrawal is unambiguously good. On the third, fees on link-outs, there is real ambiguity, and it's the one where Google could still act on its own. Below I've organized Rose's findings around those three issues.
1. What actually happened
The settlement (the "Revised Proposed Modified Injunction," or RPMI) would have made three changes that matter to anyone running Android DTC: it replaced in-Play distribution of rival stores with a web-based "Registered App Store" install flow, it removed the developer's right to put a download link inside a Play-distributed app, and it wrote in a right for Google to charge fees on link-outs.
All three are now off the table (at least from what we know today). Google reverts to the October 2024 injunction. And this is likely related to the court’s official independent opinion on the RPMI’s terms.
Judge Donato appointed Rose (MIT economist, former Deputy AAG at DOJ Antitrust) as an independent expert to judge whether the settlement served the public interest and the jury verdict, not just whether it served Epic and Google. That distinction is the spine of her report. Her starting premise is that a deal between the winning plaintiff and an adjudicated monopolist proves only that both prefer it to litigation: it can just as easily be the two of them splitting monopoly rents as a protected duopoly. So she gave Epic's endorsement little weight and leaned on the non-settling developers and amici. Her overall verdict was that three sets of changes "remain in tension with the purpose of the remedy and do not appear to be supported by evidence the Court requested."
Here is what those findings mean, issue by issue.
2. Store distribution: the biggest win, and Rose's sharpest critique
This is where the settlement tried to give away the most, and where the withdrawal matters most.
Under the original injunction, Google must carry rival app stores inside Google Play, and give them catalog access so they can offer Play's app catalog on day one. The settlement would have scrapped in-Play distribution and replaced it with a "Registered App Store" program: a streamlined, but still web-based, sideload install flow.
Rose was harshest here, and the reasoning is the whole game. She runs everything through one lens: actual access versus potential access**.** In-Play distribution gives rival stores actual access to the users Google spent years training to look only at Play. The Registered App Store program only reduces sideloading friction, which is potential access: it helps convert users who already went looking for a rival store, but does nothing for the far larger installed base that never thinks to look.
Put bluntly: the court's own economist told the judge the settlement's headline remedy was weaker than the order it would replace. With the withdrawal, in-Play distribution and catalog access survive, live from July 22. The stronger remedy won.
3. Link-out rights: restored
The original injunction bars Google from stopping a developer putting a download link inside its app to point users to an outside app or store. The settlement would have deleted that right for Play-distributed apps.
Rose called removing it a "costly unforced error." An in-app link, she notes, is "the most direct and lowest-cost channel" a developer has to reach its own existing users, and it becomes more important, not less, if distribution moves to web installation. She also flagged the contradiction in Google's security rationale: you can't argue web installation via an in-app link is too dangerous to allow while simultaneously proposing a settlement whose centerpiece (the Registered App Store program) is built on web installation being safe.
With the settlement withdrawn, the download-link right stands. No change required on your side.
4. Fees on link-outs: the unresolved one, and where Google could move unilaterally
This is where developers should temper the celebration. Two things are both true: the settlement's explicit fee right is dead, and the fee question is unresolved, at least according to Prof Rose.
Rose’s take is that, in her view, "the current injunction is silent on the fees Google may charge on link-outs and downloads." Silence cuts both ways. It means the settlement no longer hands Google a court-blessed right to tax link-outs. It also means the original injunction doesn't obviously forbid Google from trying. That is the gap in which Google could take unilateral action: assert a fee under the restored order and dare the ecosystem to challenge it.
There is a live argument the other way. Developer Benjamin Simon (founder of the Down Dog app, and an amicus in the case) contends the injunction's anti-tying provision already bars these fees, and points out Google charged nothing when it began complying in October 2025, then announced the fee schedule six weeks later once settlement talks were underway. Rose takes the narrower view that the injunction simply doesn't address fees. That disagreement is now the open front, and it hasn't been adjudicated.
This is the main thing we’ll be listening for in court, and Rose's recommendation is the useful part for us in her findings. She argues any fee Google imposes on link-outs should be capped at Google's demonstrated actual cost of a service it in fact provides, with the burden on Google to justify it, and that the Court should resolve on the record whether the injunction constrains fees at all.
Her economic case is one I've made in cruder form: a fee set at or near the full Play service fee "reproduces the prohibited conduct by price." Google can accomplish through a prohibitive charge what it is not allowed to do by an outright ban - as Apple got in trouble for doing in May 2025.
So the honest position is: the fee fight might not be over. It remains to be seen whether Google tries to assert link-out fees under the now-restored original injunction, and if so how hard it fights for them. Rose has handed the judge a ready-made principle (cost-based cap, burden on Google) for whenever that fight arrives.
5. What to do
- US, today: nothing changes, and your position got stronger. Keep your zero-fee direct link-outs. Google's own US page still says it is "not assessing" link-out fees, the download-link right is intact, and in-Play distribution is back. Do not pre-comply with a fee that does not exist.
- Start exploring alternative distribution. This is major - and developers should immediately start exploring
- Watch for Google testing the silence. The one thing the withdrawal did not do is definitively bar link-out fees. If Google moves to assert them, that is the moment the Rose cost-cap argument gets its hearing. Stay alert.
- Non-US: unchanged. Outside the US there is no Donato injunction, so the service fee on external links can still bite UK and EU in-app webshop flows. Model those economics before you lean on in-app links as your primary onboarding path there.
- Own the relationship regardless. Out-of-app DTC, email, community, web, stays at roughly all-in payment cost (~5%) and untaxed no matter how the fee question resolves. The in-app link-out is a strong top-of-funnel tool, but the durable economics live outside the app.
The honest caveat, and what's next
The link between Rose's report and the withdrawal is my inference, not a stated fact: neither party said "we quit because of the Rose declaration." But the timing (July 12 filing, July 15 withdrawal, on the eve of a hearing before a judge who was already openly skeptical) is a strong signal. And "settlement withdrawn" is genuinely good news. Just don't mistake it for "the fee question is closed." It isn't.
And it isn't quite over procedurally either: we will still be in court tomorrow, as the parties turn to Google's ongoing compliance with the court order. I'll be there and posting updates.
More as this develops. Reach out if you want your link-out economics modeled by market.
