Meta’s new AI app has delivered a startling opening number. In a comparison limited to iPhones in the United States and Canada, market intelligence firm Apptopia estimates that Muse drew 1.8 million downloads in its first 12 days, against 1.3 million for ChatGPT at the same stage of its mobile launch. The result is a strong signal for Meta’s distribution machine. It is not yet a verdict on which AI assistant people will keep using.
The investment question is now more precise: can Meta convert an early rush of curiosity into repeat use and subscriptions, or is Muse borrowing attention from Facebook, Instagram and WhatsApp without building a durable new business?
What changed
TechCrunch reported Apptopia’s estimates on September 21. The comparison matches geography, platform and launch age: iOS downloads in the U.S. and Canada during each app’s first 12 days. Muse’s estimated lead is 500,000 downloads, or about 38% over ChatGPT’s early iOS figure. That arithmetic is meaningful only inside this narrow comparison. It does not compare the apps’ total users today, lifetime downloads or revenue.
Meta introduced Muse on September 8 as a personal AI agent that can plan and act for users. Its launch announcement says the product is rolling out in the United States on iOS, Android and the web, with access through WhatsApp. Meta says most use is free, with subscriptions for more. The company has not disclosed its own early Muse adoption count.
Why it matters
Distribution has become the immediate battleground
An AI assistant needs repeated opportunities to help. Meta can put Muse near people who already use its communication and social apps; that may lower the cost of getting a first try. Apptopia estimates that more than 95% of Muse users also use Facebook and 63% use Instagram. Those are audience overlaps, not proof that either app drove the download.
The human test is simple. An app that can arrange travel or manage an inbox may save time, but it also asks for access to personal data and trust in automated actions. A download can reflect curiosity. Permission to connect services and the decision to return next week demand much more confidence.
The measurable market result is still ahead
For Meta investors, the potential prize is a new paid product and more time spent inside a wider app ecosystem. Yet the public data here measure installs, not paying users, retention, cost to serve or margin. There is no basis for translating the first 12 days into an earnings forecast. OpenAI, meanwhile, has an established ChatGPT user base that this launch-window comparison does not measure.
The chart keeps that distinction visible: it compares only the two verified Apptopia install estimates under the same iOS and geography rules. It does not award a market-share crown.
The case
Meta has a plausible route from launch attention to habit. Muse is available as a standalone app and can meet users in WhatsApp, while Meta’s existing services already supply daily interaction. If users find an agent reliably useful for recurring tasks, that distribution could translate into lower acquisition costs and a large subscription funnel.
The early data show that people are willing to try the product despite intense competition. A roughly 38% lead in the matched iOS launch-window estimate is too large to dismiss as a rounding artifact. It puts pressure on competitors to defend the everyday consumer relationship, not only model benchmarks.
The other view
Launches are not steady states. A heavily promoted new app can post a large first-week download count and then lose users once novelty fades. The Muse and ChatGPT launches also happened in different years, against different levels of public AI awareness and different promotional conditions. Matching platform and geography improves the comparison; it cannot erase those differences.
Apptopia’s figures are estimates produced outside Meta and OpenAI. Meta has not confirmed the adoption numbers. More important, installs alone reveal neither how often people return nor whether they pay. A user who tries Muse once through a Facebook prompt has not necessarily switched from ChatGPT.
What to watch next
First, look for independent estimates of weekly and monthly active users after the launch surge, ideally with a consistent method and geography. Retention at 30 and 90 days would tell more than another download milestone.
Second, watch for Meta to report paid Muse subscribers, pricing take-up or revenue. An explicit figure would turn an attention story into a measurable business story. Without it, subscription potential remains a hypothesis.
Third, observe whether Muse usage persists outside Meta’s own apps. If standalone and connected-service activity grows, it would suggest genuine assistant demand. If use is concentrated in cross-promoted surfaces, the strategic benefit may still be real, but the economics will be harder to separate.
Decision in 30 seconds
Muse’s estimated early iOS downloads in the U.S. and Canada beat ChatGPT’s matched mobile launch window, 1.8 million to 1.3 million. That is evidence of launch demand, not proof of lasting leadership. The next decision point is retention and paid conversion. Treat the first 12 days as a distribution test that Meta passed, with the business model still unproven.
Jasper Kellan is editor of The Perspective Log. This article is general information, not personal investment, tax or legal advice. Published September 22, 2026.
Primary sources
- Meta — Muse launch announcement
- TechCrunch — Apptopia’s early app estimates and methodology
