Meta is moving quickly from introducing Muse to explaining how the personal AI agent could make money. During its third-quarter earnings call, CEO Mark Zuckerberg said Muse has passed 10 million users following an early-September rollout in the US and Canada, and described the product as having reached “clear product-market fit.” Meta’s stated next step is scaling: it wants Muse to serve billions of people.

That ambition is not simply about adding another chatbot to the company’s apps. Meta is framing Muse as a commerce layer, a subscription product and a key part of the pitch for its AI glasses. Zuckerberg said the company is already developing ways to monetize the agent, and projected that Muse could eventually deliver more revenue per user than Facebook and Instagram.

It is a striking target, especially because Meta says Muse is costlier to operate than its existing services. The company’s proposed answer is a familiar blend of transaction-related business fees and paid tiers for people and businesses that use the agent more intensively. What remains unresolved is what features will stay broadly accessible, what will sit behind a subscription, and how much value the agent must provide before people—or merchants—will accept another fee in the purchasing process.

The numbers Meta has shared so far

Muse is currently limited to the US and Canada, yet Zuckerberg said more than 10 million people have started using it since its rollout in early September. “Started using” is an important qualification: it indicates meaningful early reach, but does not by itself reveal how often those people return, how many complete a purchase, or how much of Muse activity goes beyond shopping.

Meta nevertheless sees shopping as Muse’s most common use case. Zuckerberg said the average purchase made with the agent is about $200. That figure is central to the company’s commercial argument. A personal agent that can help a person move from intent to transaction offers a potential way to make money that differs from traditional advertising: a business could be charged when the system helps facilitate a sale.

Meta also intends to add AI subscriptions for high-volume individual users and business owners. The company has not laid out the price, feature list or exact threshold that would distinguish heavier use. It also has not said whether every transaction will create a fee, what form its commerce tools will take, or when those tools will arrive. For now, these are directions described by company leadership rather than a detailed consumer pricing plan.

What “product-market fit” means in this context

Product-market fit is business shorthand for a product finding a sufficiently strong match with a real user need. It does not mean a product is finished, universally available or proven to be highly profitable. In Meta’s usage here, the phrase signals confidence that Muse’s early audience is showing enough demand for the company to shift attention toward growth and operations.

For users, that operational phase may mean a wider set of tasks, more business integrations and more places to access the agent. For Meta, it means solving a harder question than launch-day curiosity: whether demand can scale while the cost of running the AI remains sustainable.

Related coverage includes Meta Positions Muse as a Commerce and Subscription Engine.

Why commerce sits at the center of the plan

Commerce gives Meta a direct path between an AI interaction and a measurable business outcome. If someone uses Muse to research, organize or make a purchase, the company can potentially build tools for merchants and collect transaction-related fees. Zuckerberg’s comments suggest Meta believes the agent’s current shopping behavior supports that model.

This differs from an AI assistant that is mainly used for casual questions or text generation. Those interactions may still be useful, but they do not inherently establish a payment event. Shopping does. The $200 average purchase figure, as presented by Meta, is therefore less a complete portrait of Muse activity than an indication of why the company is prioritizing commerce tools.

There is also a practical tension. A personal agent needs to feel like it is acting in the user’s interests, particularly when it is helping with purchases. Meanwhile, a platform charging businesses for transactions has its own commercial incentives. The information available so far does not explain how recommendations, fees or merchant participation will be presented within Muse. Those details will matter to users deciding how much trust to place in an agent handling shopping-related tasks.

Subscriptions are Meta’s second lever

Meta’s plans are not limited to business payments. Zuckerberg said subscriptions will be incorporated for people and business owners who use Muse more heavily. In plain terms, the company appears to be considering a freemium structure: a base experience may remain available while more intensive usage or additional capabilities could require payment.

That approach fits the operating-cost issue Meta has acknowledged. AI agents can require substantially more computing resources than many conventional app interactions, and Zuckerberg said Muse is significantly more expensive to run than Meta’s other services. A subscription can help align revenue with the users who place the largest demands on the system. It can also create a more predictable income stream than transaction fees alone.

But the public details stop there. Meta has not specified the capabilities planned for paid Muse access, whether those capabilities will differ between phones and glasses, or whether a future glasses feature will require an existing AI subscription, a separate add-on, or neither. It would be premature to treat the company’s broad subscription language as a finalized package.

The bigger point is that Meta is describing two complementary models. Commerce tools seek revenue from businesses when transactions occur; subscriptions seek revenue from users and businesses that rely on Muse most heavily. Together, they would reduce the pressure to make one model carry the entire cost of an expensive AI service.

Muse is also part of the smart-glasses strategy

Meta has positioned Muse as central to making more money from its AI glasses. Zuckerberg said the company does not currently make a large profit from the hardware itself, but argued that putting Muse on the glasses could support more AI subscriptions and help sell more devices.

Meta previewed Muse on its smart glasses at Connect last month. The planned experience is conversational: users will be able to talk with the agent in a manner similar to chatting with Meta AI, check on tasks and add items to an agent-managed to-do list. Meta says Muse will arrive on its glasses next month.

That matters because an AI agent becomes more useful when it is available at the moment someone wants to capture a reminder or follow up on a task. A glasses-based version could make spoken task management a prominent use case rather than an occasional phone interaction. Still, Meta has not detailed the final feature set, and it has not confirmed whether Muse on its glasses will require a separate subscription.

Zuckerberg’s remarks do indicate that some Muse features on the glasses could be paywalled. Meta has also considered charging extra for other glasses capabilities, including a forthcoming hearing-enhancement feature. The company is therefore outlining a model in which the hardware is not necessarily the main profit center; services and capabilities used through that hardware may be increasingly important.

The revenue comparison is a long-term forecast, not a present result

Zuckerberg said Meta’s existing family of apps earns several hundred dollars per American user annually on average, and said Muse could ultimately exceed that level because of the commerce already occurring through the agent. The comparison is deliberately ambitious. It is a forecast about future revenue per user, rather than a claim that Muse has already overtaken Meta’s apps.

Revenue per user is a straightforward measure: total revenue divided by the number of users over a period. It does not equal profit. That distinction is especially important here because Meta has said Muse is expensive to operate. A high revenue-per-user figure would not automatically mean the agent generates superior margins after computing, product development and other costs are accounted for.

Likewise, an average purchase of $200 is not the same as $200 in revenue for Meta. A purchase value describes the size of a transaction. Meta’s future revenue would depend on the fee structure it builds around transactions, the volume of eligible purchases and the level of merchant adoption. Those variables have not been specified.

Meta’s thesis is clear, however: the more an agent can help users get things done—especially actions that lead to commerce—the greater its potential to monetize beyond the conventional app model. Whether that thesis holds will depend on sustained usage, user trust, merchant participation and the eventual cost of supplying the service at much larger scale.

What to watch as Muse expands

  • Availability: Muse is currently available only in the US and Canada, even as Meta speaks about reaching billions of people.
  • Monetization details: Meta has identified commerce fees and subscriptions, but it has not released a final price or full feature breakdown.
  • Glasses implementation: Muse is expected on Meta’s glasses next month, with task check-ins, to-do additions and conversational access among the described functions.
  • Paid-feature boundaries: It remains unclear which glasses features, if any, will require a Muse subscription or another paid option.
  • Repeat use versus early adoption: The 10 million-user figure demonstrates early uptake, while longer-term engagement and completed commerce activity will be more revealing measures of durable demand.

Meta’s Muse strategy lands amid a broader push to turn AI agents into products people use proactively rather than only tools they query on demand. For another recent example of how the industry is framing this category, see the research-preview rollout of Claude Duet as a proactive AI agent.

For Meta, Muse is now being asked to do three jobs at once: justify greater AI investment, create a route to commerce and subscription revenue, and give consumers another reason to use the company’s smart glasses. The early user total and shopping focus give the company material to build on. The unanswered questions—pricing, merchant tools, feature access and the economics of large-scale operation—will determine whether this “new phase” becomes a major business or simply a very costly one.