Meta's Muse app passed 10 million US downloads within 40 days of its September 8 release, making it only the 12th app to reach that number so quickly since October 2015, according to estimates Sensor Tower published on October 20, 2026. The market intelligence firm also puts Muse at 2.5 million US daily active users and finds that Meta is directing 39% of its October advertising spend to the app.

In Short

Meta's AI agent app Muse was downloaded more than 10 million times in the US in its first 40 days, a pace only 11 other apps have matched in about 11 years, according to estimates from Sensor Tower. A big reason is that Meta is spending heavily to advertise it, with nearly four of every ten Meta ad dollars this month going to Muse. More people are still using it a week or two after downloading than with ChatGPT, Grok or Claude, but the share who keep coming back drops sharply after the first day, and Sensor Tower's own report gives two different figures for the rivals' two-week number.

What Sensor Tower measured

The report, titled "Muse Races to 10mn US Downloads" and dated October 20, 2026, is a three-page research note from Sensor Tower's insights team. Sensor Tower distributed it to press the same day by email, asking that citations credit Kara Lee, senior insights analyst, and describe the company as a "market intelligence firm." None of the figures come from Meta. Meta has not published download, user or retention numbers for Muse.

That distinction matters for how the numbers can be read. According to Sensor Tower's disclaimer, its mobile app and digital advertising estimates "are based on our proprietary panel and data science models." The firm adds that they "may not accurately represent metrics for any app or advertiser, and should be used for directional trend analysis." Every download, user, retention and advertising figure below is an estimate on those terms.

The headline numbers, according to Sensor Tower:

  • Muse has accumulated 10.9 million downloads worldwide since its launch on September 8, 2026.
  • More than 10.2 million of those came from the US.
  • The US crossed 10 million downloads within the first 40 days, a window Sensor Tower defines as September 8 to October 17.
  • The US accounts for 94% of Muse's downloads.

The roughly 700,000 downloads outside the US are not broken down by country in the report. Muse is not a global product: it went live in the US on September 8 through a standalone app and WhatsApp, and Canadian availability on iPhone and the web followed on September 18, as PPC Land reported earlier this month.

Twelfth app since 2015

Sensor Tower started tracking apps in October 2015, and in that time, according to the firm, only 11 other apps reached 10 million US downloads within 40 days. The most recent one before Muse was Meta's own Threads, in July 2023.

The composition of that group is lopsided. More than 80% of the earlier apps were mobile games, according to Sensor Tower, with entertainment and social networking accounting for 9% each. Social networking means Threads. With 11 apps in the group, 9% corresponds to a single app in each category, which leaves nine games. Sensor Tower does not name them, nor the entertainment app.

By that arithmetic, Muse is the only AI assistant in the group and only the second app from outside games and entertainment, after Threads, by PPC Land's reading of the percentages. It was not the fastest even within Meta's own portfolio. PPC Land's earlier report on Sensor Tower data noted that Threads had collected about 14 million US downloads in its first 22 days alone, a level Muse had not reached even by day 40.

How the AI rivals compare

The more direct comparison in the report is with other AI assistants. According to Sensor Tower, ChatGPT, Grok and Claude collected 3.7 million, 500,000 and 200,000 US downloads respectively in their first 40 days. Muse's 10.2 million is therefore nearly three times ChatGPT's figure for the same window and around 50 times Claude's.

Time to 10 million US downloads shows the same gap, according to the firm:

  • Muse: 40 days
  • ChatGPT: 123 days
  • Grok: 211 days
  • Claude: 673 days

These numbers extend the series Sensor Tower published on September 30. At that point, the firm put the time to 5 million US downloads at 22 days for Muse (then a projection), 56 days for ChatGPT, 103 days for Grok and 492 days for Claude. On both thresholds Muse sits well ahead, though Grok and Claude were initially available only on iOS, a detail the September note flagged and the October note does not repeat.

ChatGPT's early growth looked different in shape. According to Sensor Tower, ChatGPT reached about 15 million global downloads in its first 40 days, with the US, Japan, Brazil and Germany accounting for 25%, 9%, 6% and 5% of them. That 25% US share of roughly 15 million is consistent with the 3.7 million US figure, by PPC Land's calculation. Muse's growth, by contrast, is almost entirely domestic.

The report also notes a current-period contrast. ChatGPT's US downloads in the past 30 days declined 25% compared with the prior 30-day period, according to Sensor Tower. The report does not state the exact dates of either window, and it does not attribute the decline to Muse. It comes on top of a broader shift PPC Land has tracked: Comscore data showed ChatGPT's share of US AI prompt volume falling from 70% in January to 50% in June 2026, as Gemini and Claude gained ground.

Daily users and the 15% weekly gain

Downloads measure installs, not use. On use, Sensor Tower estimates that Muse reached 2.5 million US daily active users for the first time on Friday, October 16. Daily active users rose 15% week over week across the seven days from October 11 to October 17, according to the firm.

Sensor Tower does not publish monthly active users for Muse, and it does not say how it counts a user who interacts with Muse through WhatsApp rather than the standalone app. Since the download figures are app-store estimates, WhatsApp usage may fall outside both measures entirely. The report does not address this.

As a rough indicator, 2.5 million daily users against 10.2 million US downloads works out to about one in four, by PPC Land's calculation. That ratio is imprecise: downloads count reinstalls and multiple devices, and the daily figure was a single-day peak. It does give some sense of scale, though. Meta has not disclosed comparable data.

Retention: higher than rivals, but steeper decay

Retention is where the report makes its strongest claim and also where its numbers are least consistent.

According to Sensor Tower, US users who installed Muse in September 2026 averaged a 30% retention rate after seven days and 23% after 14 days. In other words, nearly a quarter of September installers were still returning about two weeks later.

The comparison figures for ChatGPT, Grok and Claude appear twice in the report, and they do not match. The first page gives the peer average as 19% at seven days and 17% at 14 days. The second page gives 19% and 16% for the same comparison. Sensor Tower does not explain the difference, and PPC Land could not determine which figure is correct. Either way, Muse is ahead at both intervals by the firm's own estimates.

The report also contains a figure that complicates the headline. On day one, 73% of September installers returned to Muse, according to Sensor Tower. By day 14 that had fallen to 23%, a decay the firm describes as "nearly 50 percentage points" (the two figures given are exactly 50 points apart). Over the same period, ChatGPT, Claude and Grok saw retention decay of about 12 points on average.

Taken together, those numbers suggest that Muse attracted far more first-day returns than its rivals, then lost them much faster. Working backwards from Sensor Tower's figures, a 12-point decay ending at 16% or 17% implies a peer day-one rate somewhere around 28% or 29%, against Muse's 73%, by PPC Land's calculation. Muse still finishes the two weeks ahead, but the gap at day 14 is roughly six or seven points, not the 44 or so points at day one. That shape is consistent with heavy curiosity-driven installs from advertising, followed by churn among users who tried the app once. Sensor Tower does not characterize it that way.

One further wording issue: a sentence in the retention section says "Meta is currently outpacing AI competitors in its retention," where the context indicates Muse rather than Meta's apps as a whole.

Advertising behind the growth

Sensor Tower is explicit that Meta's own advertising probably plays a role. The report says Muse "is potentially experiencing outsized gains due to the advertising push by Meta," and that Meta ranks within the top 10 US advertisers by spend month to date in October 2026.

The Muse-specific figures, according to Sensor Tower:

  • Muse ranks as the 21st largest brand in the US by advertising impressions month to date in October 2026, up from 73rd in September.
  • Meta is allocating 39% of its advertising spend to Muse month to date in October, up from 23% in September.

That ramp was already visible a month ago. Sensor Tower's September note, as PPC Land reported on October 4, estimated that Meta ramped Muse advertising spend on September 16, when daily downloads rose 73% day over day, and that Muse received up to 50% of Meta's daily house ad impressions between September 14 and 27. It also put Muse among the top 15 US brands by impressions for the week of September 21 to 27.

The 73rd-place September ranking in the October 20 note does not contradict that weekly figure, though the two are hard to reconcile without more detail. A monthly ranking includes the week before launch and the first week after it, when Meta had not yet ramped spending. Sensor Tower does not reconcile the two in the new note.

The 39% share is the more notable number for advertisers. It means that, by Sensor Tower's estimate, close to two of every five dollars Meta spends on its own advertising this month are going to a single product launched six weeks ago. The report does not say which channels carry that spend, whether it includes placements in Meta's own apps, or what the absolute amounts are.

App store rankings

Muse ranked first on the US Apple App Store on Monday, October 19, 2026, according to Sensor Tower. It held the top spot for the past 30 days, except on Saturday, October 17, when the Kick Streaming app briefly took first place overall and Muse dropped to second. (The report's phrasing, "drop one spot (two overall)," is ambiguous; second overall is the reading consistent with the rest of the sentence.)

On the US Google Play Store, Muse ranked as the top app for 30 consecutive days, from September 19 to October 19, 2026, according to the firm. The September note had already recorded Muse at the top of both US stores by September 30.

Why it matters for marketers

For the advertising market, the numbers carry two separate signals.

The first is about attention. Muse is an agentic AI product that browses merchant websites and buys on a user's behalf, and its user base is now large enough to register in third-party traffic data. DataDome reported on October 15 that Muse accounted for 33% of agentic browser requests on its network in September, after only nine days of measurement. Profound found that Muse drew four of its six shopping picks from two websites in a small test. Retailers have reacted in different ways: Amazon cut Muse off from buying on Amazon.com as of September 20. An app with 2.5 million US daily users and the top App Store ranking is, for merchants, a meaningful source of agent-driven visits that do not look like traditional search or social referrals.

The second signal concerns Meta's own priorities. Meta has also opened Muse to small businesses, with access to ad accounts and 15 business tools, and Meta has not said whether Muse will carry advertising. An analysis covered by PPC Land argued that Muse may never carry ads yet still shape Instagram ads through the merchant data it gathers. Putting 39% of the company's own advertising budget behind Muse is a measurable sign of how much weight Meta places on the product, whatever its eventual business model.

There is also a measurement caution worth stating. Sensor Tower's estimates come from a panel and models, not from app store or Meta data, and the firm itself describes them as directional. Meta's existing apps have not been growing on the same measure: Sensor Tower estimated in August that Facebook and Instagram downloads fell an average of 5% year over year in the second quarter of 2026. Muse's figures arrive from the same source, with the same caveats.

What remains unknown is whether the retention advantage survives once Meta's advertising support slows. The steep drop between day one and day 14 suggests a large share of installs came from people who tried Muse and left. Sensor Tower's September cohort covers installs made while spending was ramping. October cohorts, with 39% of Meta's spend behind them, will show whether heavier promotion brings in users who stay.

Timeline

Summary

Who: Sensor Tower, a market intelligence firm, published estimates about Meta's Muse AI agent app, comparing it with ChatGPT, Grok, Claude and Threads.

What: According to Sensor Tower, Muse reached 10.2 million US downloads (10.9 million worldwide) within 40 days, the 12th app to do so since October 2015. It hit 2.5 million US daily active users on October 16, retained 30% of September installers after seven days and 23% after 14, and received 39% of Meta's October advertising spend. The report gives two different 14-day retention averages for rivals (17% and 16%), and Muse's retention fell 50 points between day one and day 14 against about 12 points for rivals.

When: Sensor Tower published the note on October 20, 2026. The download window runs from Muse's September 8 launch to October 17.

Where: The download, user, retention and ranking figures cover the US, which accounts for 94% of Muse's downloads. Muse is also available in Canada.

Why: The figures show how quickly a heavily promoted AI agent can build an audience, how much of Meta's advertising budget is going to the product, and how many agent-driven users merchants and publishers may now see, while leaving open whether users stay once that promotion eases.