Meta: 'Muse' App Surpasses 10 Million Users, with Shopping Leading the Way

During its third-quarter earnings call, Meta announced that its artificial intelligence–based application, known as "Muse," has surpassed the 10 million user mark — a figure consistent with estimates published by market analytics firm Sensor Tower earlier that same week. What stood out in the company's remarks, however, wasn't the size of the user base itself, but rather how people are using the app and how much money they're spending through it.

Usage Rates Exceeding Expectations

According to CEO Mark Zuckerberg, roughly 25% of people who complete the app's setup process use it on a daily basis — a figure that contributes to improved user retention rates, the metric tech companies rely on to gauge the success of any new product.

More significantly, over a quarter of active users who employed the AI assistant to complete purchases recorded an average purchase value exceeding $200 — a number industry insiders described as surprising, as it far outpaces the average basket size across many traditional e-commerce channels.

Zuckerberg cited testimony from financial technology company Plaid, which provides the connective layer between payment applications and customers' bank accounts, and which Muse relies on to enable agentic shopping. He said the company informed Meta that it had never seen a comparable level of activity on any other service before.

Beyond Shopping: Productivity and Security Confidence

Beyond commerce, other popular uses of the app relate to personal productivity, such as connecting the AI assistant to email or a personal calendar. Zuckerberg views users' willingness to grant these sensitive permissions as a sign of confidence in the product's security architecture, which includes a component called "Muse Secure VM" dedicated to protecting data and privacy. He also referenced a forthcoming version called "Muse Confidential VM" that will prevent even Meta itself from accessing user content.

From Product-Market Fit to Scale and Profitability

With the app reaching what Zuckerberg described as "product-market fit," Meta's priorities are shifting to three areas:

• Scaling the AI assistant to serve a far larger user base.
• Increasing the revenue it generates.
• Improving the cost efficiency of the computing power required to run it.

The most notable expectation here is the company's belief that average revenue per user on Muse will eventually exceed what its current family of apps — Facebook, Instagram, WhatsApp, and Messenger — generates, which currently sits at around "several hundred dollars per year" per user, according to Zuckerberg's statements.

The Monetization Model

Meta has previously explained that its plan involves keeping the app free for most users while retaining a percentage of commerce transactions conducted through the platform. Heavy users who want access to the latest models or larger token allowances will have to pay.

On the cost front, Zuckerberg acknowledged that running Muse is very expensive in the near term from a computing standpoint, but he projected more than a tenfold improvement in cost efficiency over the coming year, which would make the app profitable once revenue reaches sufficient scale. He added that the company is building computing capacity sufficient to serve more than a billion people within the next few years.

A Pivotal Role in the Smart Glasses Strategy

The AI assistant will also play a central role in Meta's smart glasses strategy. Reality Labs, the division responsible for glasses, virtual reality, and the metaverse, continues to post losses in the billions, though those losses are expected to narrow from $19.1 billion in 2025 to around $15 billion this year.

The 'Muse Network': From Single Player to Collective Interaction

Zuckerberg hinted at a new direction he called the "Muse network" — a shift from an individual experience resembling a "single-player game" to an environment where your AI assistant can interact with those of your friends and family members. In a commercial context, this means a purchase won't just be a matter of navigating a store's website, but a direct conversation with that store's own AI agent.

What Does This Mean for Businesses?

These figures show that agent-driven shopping has moved from the experimental phase into actual behavior with substantial purchase volumes. For online retailers and service providers, this raises a practical question: are your systems ready to handle an AI agent that browses, compares, and buys on a customer's behalf? The answer to that question may well become a decisive factor in digital competitiveness over the next two years.

✦ بقلم فريق دروب أيديا