Mark Zuckerberg is positioning Meta’s new AI agent Muse around a business model that differs from the subscription-heavy approach used by many AI competitors. Speaking at Meta Connect, Zuckerberg said Meta is making Muse available for a huge amount of usage at no cost because the company expects the product to help users make or save money, with Meta eventually taking a small share of transactions generated through the platform.

Meta Is Betting on Free Access to Drive Scale

Muse launched on September 8 with a free tier alongside paid plans priced at $20 and $100 per month for users seeking greater capacity. Meta has said the vast majority of users should be able to accomplish what they need through the free version, while Zuckerberg described the free offering as a way to make advanced AI broadly accessible.

The approach is important because AI agents require substantial computing resources, making unlimited free usage difficult to sustain under a conventional software model. Instead, Meta is effectively treating free access as a mechanism for building a large user base and encouraging people to conduct more activities through Muse. The platform can perform tasks including research, scheduling, shopping and other online transactions, giving Meta a potential economic role beyond simply providing an AI chatbot.

Transaction Fees Could Change Meta’s AI Economics

Zuckerberg’s comments point to a commerce-based monetization model in which Meta would take a small fee from the economic value created by Muse. That could include transactions completed through connected merchants, purchases made through the agent or activities where Muse helps users operate a business and generate revenue.

The strategy would represent a significant shift for Meta, whose core business remains heavily dependent on advertising. In the second quarter of 2026, Meta generated $60.8 billion in revenue, including $59.4 billion from advertising. At the same time, the company spent $31.1 billion on capital expenditures during the quarter and expects full-year capital expenditure of $130 billion to $145 billion, underscoring the scale of investment required to support AI infrastructure.

Muse Needs an Ecosystem, Not Just Users

The transaction model also depends on businesses participating in Muse’s ecosystem. Meta has been expanding integrations with payment providers and merchants, including PayPal, Shopify and other commercial partners, while enabling Muse to interact with external services. The more transactions Muse can facilitate, the greater the potential revenue opportunity for Meta without requiring users to pay directly.

Early adoption provides an important test. Muse reached 2.8 million global downloads during its first 12 days and quickly became the top free iPhone application in the United States. However, downloads alone do not establish a sustainable business model. Meta will need to demonstrate that users return frequently, conduct meaningful transactions and trust Muse with increasingly important tasks.

Going forward, investors will likely watch transaction volumes, merchant integrations, paid-tier conversion and the economics of serving free users. Competition from other AI agents and resistance from major platforms could limit the model, while successful adoption could give Meta a new source of revenue that is less dependent on advertising. The central question is whether Meta can turn Muse from a widely used free AI assistant into an infrastructure layer through which economic activity increasingly flows.