Meta Muse Has Profitability Problem Despite Major Download Success

On September 8, 2026, Meta introduced "Muse," a personal AI agent that sends emails, books travel, and completes purchases – all based on a single instruction. The service is available exclusively to users in the US and Canada and recorded 2.5 million downloads in North America within thirteen days. This surpassed ChatGPT's download record and climbed to number one among free app downloads in the United States, as Bloomberg reported on September 21, 2026. Meta's stock subsequently gained 27 percent.

However, while users are currently testing the service virtually free – Meta grants up to 100 million tokens weekly – the company lacks a sustainable business model. Chief Executive Mark Zuckerberg outlined the direction on September 8, 2026 on Threads: "The business model we envision is to take a very small percentage of transaction volume as a fee." The fees should come from businesses, not end customers.

Cost Structure Complicates Profitability

The profitability calculation is challenging. "Making it profitable will be hard, even if Meta Muse can establish itself as an indispensable service," according to a current analysis from October 9, 2026. The decisive test is "whether accepted revenues or savings exceed agent fees, transaction costs, review time, errors, and the value of granted access."

Meta remains "free for now, while the company seeks a business model around its personal AI agent," CNBC wrote on October 7, 2026. The challenge lies not only in technical capability but in economic viability: Muse works continuously in the background – even when the app is closed – and only requests confirmation for final decisions like sending emails or making payments. The associated computational effort costs money, while revenue is currently lacking.

Capital Expenditures and Cash Flow Under Scrutiny

Meta's massive spending programme for artificial intelligence initially met with investor scepticism, particularly because of its concentration on AI compared to other technology companies like Microsoft. Muse, however, "sparked an entirely new discussion about AI use cases," Business Insider reported on September 23, 2026.

In a Reddit discussion on October 1, 2026, it was noted that "Meta has spent heavily on AI expansion," with the company generating over 40 billion US dollars in free cash flow over the past twelve months, though in the cited quarter it was less than 2 billion US dollars. The question of whether AI investments will pay off therefore remains open.

Competitors with Different Approaches

Muse differs from closed competitors like ChatGPT, Claude, Gemini, and Grok through the combination of a closed consumer app with open-source technology for developers. OpenAI's "GPT-6 Astra" handles diverse professional tasks quickly, but does not function as a life management assistant with calendar and payment integration. Anthropic's "Claude Fable 5.1" specialises in programming, enterprise knowledge work, and long-term autonomous tasks, but does not function as a turnkey service for ordinary users without separate configuration.

The announcement of new Muse updates triggered market reactions: shares of Airbnb, Expedia, and TripAdvisor each fell around five percent last week after Meta announced the expansions.

Impact on Existing Business Models

AI agents like Muse confront companies that benefit from friction and customer inertia with challenges. Goldman Sachs created a "Consumer Inertia Basket" to monitor companies that could lose profit margins to AI agents – including fitness studios with difficult-to-cancel subscriptions and financial institutions that rely on customers not switching accounts.

Muse's early success signals, according to an industry assessment from September 24, 2026, "the next phase of the AI agent economy," with 2.5 million downloads in two weeks suggesting that agent-based AI "is transitioning from a developer curiosity to mainstream consumer behaviour." This raises questions for companies about platform interoperability, data management, and delegated authority.

Data Protection and Error Tolerance

Vishal Shah, Vice President of AI Products at Meta, explained: "We cannot promise that Muse will never make a mistake, but we designed the service to protect personal information as thoroughly as possible." For purchases, Muse uses virtual one-time card numbers generated in partnership with financial platform Stripe to block data breaches. The service has memory functions that remember personal information and previous conversations to inform responses.

Zuckerberg's Strategic Positioning

Zuckerberg positioned Muse as a "personal AI agent that understands your goals and works around the clock," emphasising that the company offers "up to 100 million tokens of usage free per week, so everyone has access to personalised superintelligence." The term "superintelligence" adopts terminology that US President Donald Trump promoted as an alternative to the standard designation for artificial intelligence. Meta called Muse "the first step towards personalised superintelligence."

Whether this vision translates into a sustainable business model remains to be seen. The technical resonance is there, the economic viability is not. For investors holding or considering Meta shares, the question remains central: Can Zuckerberg turn the AI agent into a profitable business – or will Muse remain an expensive prestige project without a clear source of revenue?