Meta Platforms’ new Muse AI agent gives the company a potential route to earn money from AI through consumer transactions, extending its business beyond advertising as spending on data centers and computing infrastructure accelerates.
Muse, launched on Sept. 8, is designed to perform multi-step online tasks including web searches, price comparisons, form completion and product orders. Users must confirm payments and other sensitive actions, but the agent can handle parts of a shopping process in the background, including while a phone is locked.
The launch arrives as Meta funds an unusually large expansion of AI infrastructure. The company forecast full-year capital expenditures of $130 billion to $145 billion, a level that has put pressure on cash generation even as its advertising business continues to produce substantial revenue.
Meta reported $60.8 billion in second-quarter revenue, including $59.3 billion from advertising. Yet total costs and expenses rose 55% during the period, while operating margin declined to 31% from 43% a year earlier. Operating cash flow was $31.8 billion, but capital expenditures and lease principal payments exceeded $31 billion, leaving free cash flow of $784 million.
Muse offers Meta a way to connect those infrastructure investments to commerce activity. Rather than using AI mainly to improve feed recommendations, advertising targeting or content creation, the company is testing an assistant that can participate directly in the path from product discovery to checkout.
A commerce agent built to take action
Muse is part of a growing category of “agentic” AI products: software that can execute tasks across websites and services instead of merely generating answers or recommendations.
A user might ask Muse to find a particular product within a budget, compare offers from different retailers, apply available options and move through checkout. The agent’s ability to complete several actions without requiring continuous manual input distinguishes it from a conventional chatbot or search assistant.
That capability also puts Meta in the middle of a commercially valuable part of the internet. Search, product selection, payment and order confirmation have historically been controlled by retailers, search engines, marketplaces and payment processors. An AI agent that helps users navigate those steps could give Meta new ways to charge merchants, collect transaction-related fees or share in payment revenue.
Muse is available through a free tier and subscription plans priced at $20 a month and $100 a month. If 10 million users chose the $20 monthly option, that would produce $2.4 billion in annual subscription revenue before accounting for free users, higher-priced plans, refunds, platform fees or operating costs.
Subscriptions alone would remain small beside Meta’s advertising operation. The larger commercial opportunity would depend on whether Muse can become a meaningful source of shopping traffic and completed orders. Possible models include commissions on transactions, merchant payments for referral traffic and revenue-sharing arrangements connected to checkout services.
Retail platforms are setting different boundaries
The early response from online commerce platforms shows that agent-led purchasing will depend on access as much as consumer interest.
Amazon has restricted Muse from placing orders on a user’s behalf, limiting the agent’s ability to complete transactions through one of the largest online retail ecosystems. Shopify, meanwhile, has enabled Shop Pay within Muse’s transaction flow, giving the agent a route to complete purchases for participating merchants.
Those contrasting approaches create a practical challenge for Meta. A personal agent can be useful for research and price comparison without direct retailer integrations, but its value rises sharply when it can reliably carry a shopper through payment, confirmation and order tracking.
Retailers also have incentives to retain control over customer relationships, product discovery and payment data. Agent traffic could lower the visibility of sponsored placements or retailer-designed shopping experiences if users increasingly delegate browsing decisions to an AI assistant.
For Meta, Shopify’s participation offers an early example of how agent checkout could work without requiring the company to become a retailer or payment processor itself. It could instead direct customers toward merchants and share in the economics generated along the way.
Privacy commitments will face a commercial test
Meta has said Muse’s private conversations and its Secure VM, or secure virtual machine environment, are not directly connected to the company’s advertising system. The separation addresses an immediate concern: whether a user asking an AI agent to research a medical product, a financial service or a sensitive purchase could have those conversations fed into ad targeting.
The policy could also constrain Meta’s most familiar monetization playbook. Advertising remains the company’s financial engine, accounting for roughly 98% of second-quarter revenue based on its reported figures. If Muse data is walled off from advertising systems, commerce revenue would need to stand on its own rather than rely on direct targeting gains from private agent activity.
That separation may help Meta win trust for tasks involving payments and personal information. Yet it raises a separate governance issue around recommendations. When an agent chooses which products to show, which seller to prioritize or which payment option to use, users and merchants will want clarity on whether those results are ranked by price, availability, convenience, commercial arrangements or another factor.
Shares rose before Muse reached users
Meta’s stock climbed from the $570 range in late August to $741.25 by Sept. 21, a gain of nearly 30%. Shares rose 11.34% on Sept. 21 alone.
The timing makes it difficult to attribute the move to Muse. The product launched after the earlier advance had already begun, while Meta’s broader AI strategy, advertising performance and spending plans were already central to the company’s market narrative.
Muse nevertheless gives traders a more concrete consumer product through which to judge Meta’s AI spending. Data centers and specialized chips are expensive long before they produce visible revenue. A consumer agent that can eventually earn subscriptions, referral fees or payment-related income would give Meta a business case that is easier to measure than general promises about more capable AI.
Meta also has a large potential distribution base. The company reported an average of 3.6 billion daily active people across its family of apps in June 2026. Turning even a small portion of that audience into recurring Muse users or commerce customers would create a sizable business, though adoption, retailer access and payment partnerships remain the variables that will determine whether the product develops into a material revenue source.
Interested in AI-driven automation like Meta’s Muse? Explore Toobit’s AI copy trading for smarter, hands-off crypto investing.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
