Meta Says Muse Has Surpassed 10 Million Users as Shopping and Productivity Drive AI Growth

Meta says its AI app Muse has crossed 10 million users, and the company now believes the product is moving from early traction into a new phase of scale. The milestone matters because CEO Mark Zuckerberg says Muse is already showing unusually strong engagement, high-value shopping behavior and a path toward profitability.

The disclosure came during Meta’s third-quarter earnings call on Wednesday, when Zuckerberg outlined how the company is seeing people use Muse for everything from agentic shopping to personal productivity tasks. He said the app has become meaningful enough to shape Meta’s revenue expectations, its infrastructure spending and even its smart glasses roadmap.

Meta confirms Muse has passed 10 million users

Meta said Muse has now been adopted by more than 10 million people, confirming estimates that were circulating earlier in the week from market intelligence firm Sensor Tower. For a product that is still relatively new, that level of adoption gives Meta a credible foothold in the crowded consumer AI market.

The bigger story, however, is not simply the user count. Meta is emphasizing how people interact with Muse and how often they return to it. According to Zuckerberg, around a quarter of users who sign up for the app become daily users, a retention rate that suggests the service is becoming part of a regular routine rather than a novelty download.

That daily usage figure is especially important for Meta because AI applications are expensive to operate. A product that gets installed but rarely used can be a cost center. One that keeps people returning every day can justify heavy spending on compute, models and product development.

Why is shopping such a big deal for Muse?

Shopping is a big deal for Muse because it appears to be one of the clearest signs that users are willing to let the AI agent take action on their behalf. Zuckerberg said more than a quarter of Muse’s engaged users who used the app for shopping were making purchases averaging more than $200, a number he described as surprising.

That spending level suggests Muse may be doing more than answering questions or generating text. It appears to be influencing real transaction behavior, which is exactly the kind of outcome that can matter to Meta’s business model if the company takes a cut of commerce completed through the platform.

Zuckerberg said Plaid, the financial technology company that powers account connections for many payment services, told Meta it had never seen comparable activity through one of its connectors on another service before.

Meta uses Plaid’s infrastructure to support agentic shopping in Muse, allowing the AI to connect with users’ financial accounts and help complete purchases. That kind of integration puts Muse in a more advanced category than simple chatbots, since it is not just recommending products but helping facilitate transactions.

What the shopping data suggests

The shopping data points to three things Meta likely cares about most:

• Users appear comfortable enough with Muse to authorize financial connections.
• Shopping activity is producing meaningful transaction values rather than small impulse buys.
• The app may have a monetization path through commerce, not just subscriptions.

For a company that has built one of the world’s largest ad-driven platforms, that creates a familiar playbook. Meta has long monetized intent and behavior at scale. Muse could become a new version of that model, but with AI systems mediating the interaction instead of feed rankings or search ads.

How people are using Muse beyond shopping

Meta says shopping is only one part of the picture. Zuckerberg said another popular use case is productivity, especially when people connect Muse to their email or calendars. In practice, that means users are trusting the AI with some of their most sensitive personal workflows.

That trust is significant because AI tools tend to handle different kinds of data in very different ways. A casual chatbot may only need a prompt. A productivity assistant may need access to private messages, schedules and other personal context. Meta is betting that Muse can serve both consumer convenience and deep personal utility.

The company is also framing that usage as evidence that its privacy and security architecture is working. Zuckerberg highlighted Muse Secure VM, a feature designed to protect user data and privacy while the AI operates on connected accounts and personal information.

Zuckerberg said Meta is also preparing a future version called Muse Confidential VM, which he said will be designed so that even Meta cannot view users’ content.

That claim is notable because it signals how seriously Meta is treating privacy concerns around agentic AI. The more an AI assistant handles purchases, emails and calendars, the more users will want assurance that sensitive information is not exposed to the company itself or to unauthorized access.

How will Meta make money from Muse?

Meta expects Muse to become a meaningful revenue product, but not necessarily through direct subscriptions alone. Zuckerberg said the company’s plan is to keep the app free for most users while taking a share of commerce completed on the platform. Power users who want access to more advanced models or more tokens will likely have to pay.

That approach mirrors the broader structure of many consumer internet businesses: free access for mass adoption, then premium tiers for heavier users. But in Muse’s case, the economics may be more complex because the app is also expected to generate commerce-related revenue in addition to any subscription income.

Zuckerberg went further, saying Meta believes Muse could eventually generate more revenue per user than its core “family of apps,” which includes Facebook, Instagram, WhatsApp and Messenger. He said current annual revenue per user from those apps is in the “several hundreds of dollars” range.

If that forecast proves accurate, Muse would not just be another AI feature. It would become a major consumer product in its own right, capable of rivaling Meta’s most profitable platforms on a per-user basis.

Meta’s planned monetization mix

Based on the company’s comments, Muse’s revenue model appears to include three layers:

1. Free access for most users to encourage broad adoption.
2. A commerce take rate on purchases made through the AI agent.
3. Paid upgrades for power users who want more advanced capabilities.

That mix is designed to maximize reach while preserving optionality. It allows Meta to monetize casual users indirectly through commerce and extract direct revenue from heavier users who need more performance or capacity.

How does Muse fit into Meta’s compute strategy?

Muse fits into Meta’s compute strategy because it is currently expensive to run, but the company believes those costs can be brought down sharply. Zuckerberg told analysts that Meta is building capacity to support more than a billion people over the next few years, and he said Muse is among the workloads driving that expansion.

He also said the app is “compute-intensive” today, but Meta expects to improve cost efficiency by more than 10x over the next year. If that happens, the company believes Muse can become profitable once revenue scales up.

That is a familiar tension in AI. The products that generate the most excitement often also consume the most resources. For Meta, the challenge is to keep users engaged while making the infrastructure bill manageable enough that the economics work at scale.

What role will Muse play in Meta’s smart glasses push?

Muse is expected to play a central role in Meta’s smart glasses strategy because the company sees AI as the interface that will make wearable hardware more useful. Reality Labs, Meta’s unit responsible for smart glasses, virtual reality and the metaverse, has continued to post large losses, but those losses are forecast to narrow over time.

Meta said Reality Labs losses are expected to improve from $19.1 billion in 2025 to $15 billion this year. Even with that improvement, the division remains expensive. The logic behind Muse is that a powerful AI assistant can make hardware more compelling and potentially help Meta justify the long-term investment.

If users increasingly rely on Muse to shop, manage messages and organize their schedules, that behavior could translate naturally to a wearable device. Smart glasses are far more useful if an AI assistant can act in the background, understand context and help users complete everyday tasks hands-free.

That is where Meta’s hardware and software ambitions begin to reinforce each other. Muse may eventually serve as the software layer that makes the company’s hardware easier to adopt and more sticky to use.

What is the “Muse network” Meta is talking about?

The “Muse network” is Zuckerberg’s shorthand for a more connected version of the product, one that moves beyond a single-user experience and allows multiple AI agents to interact. He described it as turning Muse from a “single-player game” into something more collaborative, where a user’s AI can work alongside friends’ and family members’ AI systems.

That concept is still abstract, but it suggests Meta is thinking beyond individual prompts or one-on-one assistant sessions. In Zuckerberg’s vision, Muse could become a social layer for AI interaction, much like Meta’s apps turned social networking into a mass-market behavior.

Zuckerberg said the future of the product could involve an assistant that does not simply send users to a business website, but instead communicates directly with the business’s own agent.

This is an important clue to Meta’s long-term ambitions. If AI agents increasingly transact with each other, the consumer experience could shift from browsing websites to delegating tasks across connected systems. In that world, Meta would not just host an assistant; it would operate an intermediary layer for digital commerce and communication.

Why that matters for the AI industry

If Meta is right, the next wave of consumer AI may be less about isolated chatbots and more about networks of autonomous systems negotiating, purchasing and coordinating on behalf of people. That would change how businesses market products and how users interact with services online.

It would also raise new questions about standards, trust and competition. If one company’s AI becomes the gateway to shopping and personal tasks, that company could gain influence over which services users discover and which purchases they complete.

How Muse compares with Meta’s broader app portfolio

Meta’s family of apps remains the company’s financial engine, with Facebook, Instagram, WhatsApp and Messenger still generating substantial revenue through advertising and engagement. Zuckerberg’s comments suggest Muse is being positioned not as a side project, but as a future product that could rival those apps on revenue per user.

The comparison is striking because Meta is essentially arguing that a consumer AI assistant could become as commercially important as its biggest social platforms. That is a bold bet, but it fits the company’s history of scaling consumer behavior to extraordinary levels.

Muse may also be strategically useful because it gives Meta a way to participate in the broader AI economy without depending entirely on advertising. As the market for AI services matures, platforms that can convert usage into direct utility, commerce and subscriptions may have a stronger position than those that rely solely on ad clicks.

Timeline of Muse’s rise and Meta’s latest update

The public details around Muse have come together quickly, with the most recent update placing the product firmly in the spotlight during Meta’s earnings call.

What the numbers say about Meta’s AI strategy

The numbers Meta shared paint a picture of a company trying to turn an AI app into a durable consumer platform. Ten million users is enough to matter. A 25% daily-use rate points to real retention. Shopping transactions above $200 suggest commercial value. And a 10x efficiency goal shows how aggressively Meta thinks it must improve the economics.

Those figures also help explain why Zuckerberg is speaking so openly about profitability. The company is trying to convince investors that Muse is not just an experimental AI product, but one that can eventually generate material returns.

That message is especially important as Meta continues to spend heavily on AI infrastructure and on Reality Labs. The company wants the market to see those costs not as open-ended losses, but as investments in a broader ecosystem that may pay off across software, commerce and hardware.

What happens next?

The next stage for Muse will likely be defined by three questions: can Meta keep users engaged, can it reduce the cost of running the service, and can it convert agentic behavior into repeatable revenue? Those are the metrics that will determine whether Muse becomes a major business or remains a promising but expensive experiment.

For now, Meta is clearly encouraged. The company has a product with visible adoption, surprising shopping behavior, strong retention and a clear fit with its broader hardware ambitions. The challenge is that all of those strengths still need to be translated into a sustainable business model.

If Meta succeeds, Muse could become one of the company’s most important consumer products of the next decade — not just an AI app, but a platform for shopping, productivity and connected digital agents. If it fails, it will still have offered a preview of where the company thinks consumer AI is headed.

Either way, Meta is no longer talking about Muse as a side feature. It is now describing it as a product with scale, revenue potential and strategic importance across the company’s consumer and hardware businesses.