Meta Says Muse Has More Than 10 Million Users, with Shopping a Leading Activity

Preface

Meta’s AI assistant, Muse, has reportedly surpassed 10 million users, and the company says people are already relying on it for more than simple questions. During Meta’s third-quarter earnings call on Wednesday, CEO Mark Zuckerberg highlighted shopping as one of the service’s notable activities, alongside personal productivity tasks such as managing email and calendars. He also described early signs of user engagement, spending, and trust in the assistant’s security features. These developments matter because Meta is beginning to position Muse not merely as a standalone AI product, but as a potential business, a component of its smart glasses strategy, and a platform where AI agents may interact with one another. This article reviews Meta’s statements, the figures it shared, its planned monetization approach, and the infrastructure challenges that could shape Muse’s growth.

Lazy bag

Meta says Muse has attracted more than 10 million users, with about 25% of people who set it up using it daily. Shopping is a prominent activity: over a quarter of engaged users shopping through the AI agent had an average purchase size of over $200. Zuckerberg also cited productivity uses and described security features intended to protect users’ information. Meta plans to scale Muse, pursue revenue from commerce and paid advanced access, and reduce its compute costs by more than 10x in the next year. The company also sees Muse as part of its smart glasses plans and has teased a future “Muse network” connecting users’ agents.

Main Body

Meta said on Wednesday, during its third-quarter earnings call, that its AI app Muse has been adopted by more than 10 million users. The figure confirmed estimates released earlier in the week by market intelligence firm Sensor Tower. While the user total indicates the assistant has reached a substantial audience, CEO Mark Zuckerberg focused on how people are using Muse and what their activity might mean for the product’s future.

Zuckerberg said about 25% of people who set up Muse now use it daily, a level of activity that he said is helping improve retention. He also pointed to shopping as an unexpectedly significant use case. Over a quarter of Muse’s engaged users who were shopping through the AI agent had an average purchase size of over $200. Zuckerberg said this result surprised even industry insiders, suggesting that some users are already using the assistant in ways that can lead directly to purchases.

He cited Plaid, a financial technology company that connects payment apps to customers’ bank accounts, as one indication of the activity. “Plaid told us that they hadn’t seen activity like they’re seeing with their connector on any other service before,” Zuckerberg said. Muse uses Plaid’s connector to support agentic shopping. The account suggests that the assistant’s commerce activity involves more than product discovery: its connection to financial services can help enable transactions as part of the shopping experience.

Shopping is not the only area in which users are engaging with Muse. Zuckerberg said people also use it for personal productivity tasks, including connecting the AI agent to their email or calendars. Such tasks can involve personal information, making privacy and security central to whether people are willing to rely on an assistant. Zuckerberg interpreted these uses as evidence that users trusted Muse’s security setup.

That setup includes a feature called “Muse Secure VM,” which is designed to protect users’ data and privacy. Zuckerberg also mentioned a forthcoming feature called “Muse Confidential VM.” He said it will prevent even Meta from seeing users’ content. The two features were presented as part of the company’s approach to handling information as Muse takes on tasks involving users’ personal accounts and communications. Meta’s claims describe its intended protections; the call did not provide further technical details about how the features work.

Zuckerberg said Muse had achieved “product-market fit,” and that Meta’s focus would now shift toward scaling the AI agent, increasing revenue, and optimizing its compute costs. That shift reflects a transition from establishing whether users want the product to serving more people and developing sustainable business models. Meta is forecasting that Muse could eventually generate more revenue per user than the company’s “family of apps” users do today. That group includes users on Facebook, Instagram, WhatsApp, and Messenger. Zuckerberg described the current figure for those users as around “several hundreds of dollars annually.”

Meta has outlined a monetization approach that would keep Muse free for most people while retaining a percentage of commerce taking place on the platform. Users who want access to the most advanced AI models or more tokens would also have to pay. Zuckerberg said Meta believes Muse will become profitable after it scales. This plan links the company’s business prospects to both shopping activity and paid access, while leaving the basic service available to a broad audience.

The economics of operating Muse will depend in part on computing capacity and efficiency. “On the compute side, we are building out the capacity to serve more than a billion people over the next few years,” Zuckerberg told bankers and analysts on the Q3 earnings call. He acknowledged that Muse is very compute-intensive to run in the near term. Meta expects to improve its cost efficiency by more than 10x in the next year, which Zuckerberg said would enable the service to be profitable once its revenue lines are at scale. These are company expectations, rather than reported outcomes, and the relationship between lower costs, user growth, and revenue will be important to the product’s long-term prospects.

Muse is also expected to play a central role in Meta’s smart glasses strategy. The hardware and immersive-technology division, Reality Labs, is responsible for smart glasses, VR, and the metaverse, and is still losing billions. Meta expects the division’s losses to improve from $19.1 billion in 2025 to $15 billion this year. The figures place the company’s AI ambitions alongside a broader effort to develop hardware products and services. If Muse becomes closely integrated with smart glasses, the assistant could extend beyond an app and become part of how users interact with Meta’s devices.

Zuckerberg also teased a concept he called the “Muse network.” He described it as a move from a “single-player game” toward a system in which “your Muses can work with your friends’ and your families’ [Muses].” The idea suggests that agents could eventually coordinate with one another rather than acting only for an individual user. Zuckerberg connected that possibility to commerce as well: “And if you want to buy something, it’s not just navigating to that business’s website; it’s talking to their agent,” he added.

The concept points toward a broader vision in which businesses, families, and individuals may each use AI agents that communicate to help complete tasks or purchases. However, the earnings call offered a description of the direction Meta is considering, not a detailed product specification or launch schedule. Questions about user consent, privacy, reliability, and how agents represent their owners would need to be addressed as such systems develop.

Taken together, Meta’s remarks present Muse as an AI product with early adoption, measurable engagement, and a notable commerce use case. The company is also emphasizing productivity, security, future agent-to-agent interactions, and integration with smart glasses. At the same time, its plans depend on scaling the service, improving compute efficiency, and converting activity into revenue. The figures and forecasts shared by Zuckerberg provide a view of Meta’s ambitions, while the product’s lasting performance will depend on whether users continue to trust and use it and whether its costs can be brought in line with its commercial potential.

Key Insights Table