How is Meta Muse driving high-value agentic commerce?
Meta Muse is transitioning from a simple AI assistant to a powerful engine for agentic commerce, characterized by high-value transactions. During the Q3 earnings call, CEO Mark Zuckerberg revealed that more than a quarter of engaged users who utilize the AI agent for shopping have an average purchase size exceeding $200. This transaction volume has surprised even industry observers and internal teams.
The technical backbone of this shopping experience relies on a deep integration with financial technology. Meta utilizes a connector from Plaid, a company that links payment applications directly to customer bank accounts, to facilitate seamless, agentic transactions. According to Zuckerberg, Plaid reported that they had not observed activity similar to what they are seeing with their Muse connector on any other service previously.
The mechanics of agentic shopping
Agentic commerce differs from traditional e-commerce because the AI does not merely suggest products; it acts as an intermediary that can navigate the purchasing process. By leveraging Plaid's infrastructure, Muse can interact with the financial layer of a user's life to execute orders. This reduces friction in the buying cycle, moving from a search-and-click model to a conversational, intent-based model where the agent handles the logistical hurdles of the transaction.
What are the primary use cases for Muse users?
While shopping is a dominant driver of value, Muse users are increasingly employing the AI for complex personal productivity tasks. Zuckerberg highlighted that users are frequently connecting the AI agent to sensitive personal data streams, such as email accounts and digital calendars, to manage their daily schedules and information flows.
The adoption of these productivity features suggests a high level of user trust in Meta's ability to manage personal information. To support this, Meta has implemented specific security architectures designed to protect user privacy while allowing the AI to function effectively. These include:
- Muse Secure VM: A current feature designed to protect user data and maintain privacy during active sessions.
- Muse Confidential VM: A forthcoming technology that Meta claims will prevent even the company itself from accessing user content.
This emphasis on privacy is critical as Meta seeks to move the app from a "single-player game" to a more collaborative environment, often referred to as the "Muse network." In this future state, a user's Muse could potentially interact with the Muses of friends, family, or even business agents to coordinate complex tasks or purchases.
How does Meta plan to monetize and scale Muse?
Meta is targeting a revenue model for Muse that could eventually surpass the per-user revenue generated by its core "family of apps," which includes Facebook, Instagram, WhatsApp, and Messenger. Currently, the annual revenue per user for those platforms stands at several hundred dollars, according to Zuckerberg.
The monetization strategy for Muse is expected to be a hybrid approach. Meta intends to keep the service free for the majority of its user base while capturing a percentage of the commerce transactions that occur through the platform. Additionally, Meta plans to introduce a paid tier for "power users" who require access to more advanced AI models or higher token limits. To achieve long-term profitability, Meta is focusing on two main pillars:
- Scaling Capacity: Building out compute infrastructure to support a potential user base of over one billion people in the coming years.
- Cost Optimization: Improving compute efficiency by more than 10x over the next year to offset the current high costs of running intensive AI models.
What role does Muse play in Meta's hardware strategy?
Muse is positioned as a central component of Meta's broader hardware ecosystem, specifically within the Reality Labs division. While Reality Labs has faced significant financial headwinds, the division's losses are projected to narrow. According to Meta, losses are expected to improve from $19.1 billion in 2025 to an estimated $15 billion this year.
The integration of Muse into smart glasses represents a strategic move to bridge the gap between AI software and wearable hardware. By embedding the agentic capabilities of Muse into glasses, Meta can provide users with a hands-free, multimodal AI experience. This alignment suggests that the future of Meta's metaverse and hardware ambitions is deeply intertwined with the success of its agentic AI software.
Frequently asked questions
How many users does Meta Muse currently have?
Meta has confirmed that Muse has surpassed 10 million users. This figure aligns with recent estimates provided by the market intelligence firm Sensor Tower, indicating rapid adoption of the AI agent since its release.
What makes Muse shopping different from traditional e-commerce?
Muse utilizes agentic commerce, meaning the AI agent can perform tasks on behalf of the user using tools like Plaid to connect to bank accounts. This allows for a more automated, conversational shopping experience rather than manual browsing.
How does Meta protect user privacy in Muse?
Meta uses a feature called "Muse Secure VM" to protect data. Furthermore, the company is developing "Muse Confidential VM," which is intended to ensure that even Meta cannot access the content handled by the AI agent.
Will Muse be a paid service?
Meta plans to keep Muse free for most users but will monetize through a percentage of commerce transactions. Power users who want advanced models or more tokens will be able to pay for a subscription.
Is Muse expected to be profitable?
Yes, Zuckerberg stated that Meta believes Muse will become profitable once it scales. The company is working to improve compute cost efficiency by more than 10x to ensure the business model is sustainable.
Key takeaways
- Meta Muse has exceeded 10 million users, confirming growth estimates from Sensor Tower.
- Engaged shoppers on Muse have an average purchase size of over $200.
- Meta plans to improve Muse's compute cost efficiency by 10x within the next year.
- The Muse network will eventually allow different AI agents to interact with each other.
The path to agentic dominance
Meta's rapid scaling of Muse represents a pivot toward a future defined by agentic interaction rather than traditional interface navigation. By successfully integrating high-value commerce with robust privacy protections like the Confidential VM, Meta is building a foundation that could redefine per-user monetization. While the high compute costs and Reality Labs' ongoing losses remain significant hurdles, the projected improvements in efficiency and the expansion into a multi-agent "Muse network" suggest a long-term strategy aimed at dominating the intersection of AI, commerce, and wearable hardware.
