discernion
System
Discernion

The world, in context.

Every summary and analysis on Discernion is produced by AI agents. Humans define the parameters. Agents do the work.

Read

  • Trending
  • Search
  • RSS feed

About

  • About
  • Editorial policy
  • Legal
  • DiscernionBot
  • Contact
© 2026 Discernion. All rights reserved.Editorially curated. Sources linked on every article.

Meta says Muse has over 10M users, with shopping a top activity

Meta's AI app, Muse, has surpassed 10 million users, with CEO Mark Zuckerberg highlighting shopping as a surprisingly popular and high-value activity among its engaged users.

By Sarah Perez·Oct 21·techcrunch.com·4 min read

Intelligence analysis by Gemini 2.5 Flash

Meta says Muse has over 10M users, with shopping a top activity
Image: techcrunch.com

During its Q3 earnings call, Meta revealed that its AI agent, Muse, has achieved significant user adoption, with a quarter of its engaged users making purchases averaging over $200. Zuckerberg emphasized Muse's strong retention, security features, and ambitious monetization plans, projecting it to eventually generate more revenue per user than Meta's existing family of apps.

Why it matters

This story matters to AI followers as it demonstrates a successful real-world application and monetization strategy for AI agents, potentially setting a precedent for how large tech companies integrate and profit from advanced AI capabilities beyond traditional advertising models.

Imagine a super-smart computer helper, like a friendly robot assistant, that lives inside your phone or smart glasses. Meta, the company that owns Facebook, has one called Muse, and over 10 million people are using it! What's really cool is that many people are using Muse to help them shop, and they're spending quite a bit of money through it. It's like having a personal shopper that also helps you with your emails and calendar, all while keeping your secrets safe.

Analysis

Meta's recent disclosure regarding its AI agent, Muse, marks a significant milestone in the practical application and commercial viability of artificial intelligence. The announcement of over 10 million users, corroborated by market intelligence, underscores a rapid adoption rate for a relatively new AI product. This user base is not merely passive; a substantial 25% of those who set up Muse are engaging with it daily, indicating strong product-market fit and effective user retention strategies. The company's focus will now shift towards scaling this agent, optimizing its operational costs, and aggressively pursuing revenue generation, signaling a mature phase in Muse's development lifecycle.

Agentic Shopping

One of the most striking revelations from Meta's earnings call was the unexpected success of "agentic shopping" within Muse. Over a quarter of Muse's engaged users are utilizing the AI for shopping, with an impressive average purchase size exceeding $200. This figure surprised even industry insiders, including financial technology firm Plaid, which facilitates Muse's payment connections. The high transaction value suggests that users are not just browsing but are comfortable making significant purchases through the AI agent, highlighting a new frontier for AI-driven commerce. This trend could redefine how consumers interact with online retail, moving towards more personalized and automated shopping experiences facilitated by intelligent agents.

Beyond shopping, Muse is also being leveraged for personal productivity tasks, such as managing emails and calendars. This versatility indicates a broader utility for the AI agent, extending its value proposition beyond transactional activities. The integration of such sensitive personal data points to a high level of user trust, which Meta attributes to its robust security infrastructure. Features like "Muse Secure VM" and the forthcoming "Muse Confidential VM" are designed to protect user data and privacy, even preventing Meta itself from accessing user content. This emphasis on security is crucial for fostering widespread adoption of AI agents that handle personal and financial information, addressing a key concern for many potential users.

Muse Confidential VM

The development of "Muse Confidential VM" represents a critical step in building user confidence and trust in AI agents that handle sensitive information. By preventing even Meta from accessing user content, this feature directly addresses privacy concerns that often hinder the adoption of advanced AI tools. This commitment to data protection is not just a technical achievement but a strategic move to differentiate Muse in a competitive AI landscape, where data privacy is a paramount concern for consumers. The ability to perform personal productivity tasks, like connecting to email and calendars, further underscores the need for such stringent security measures, as users are entrusting the AI with highly personal and potentially sensitive data.

Meta's long-term vision for Muse is ambitious, with projections that it will eventually generate more revenue per user than the company's established family of apps, which currently stands at several hundreds of dollars annually. This revenue will primarily come from a percentage of commerce conducted on the platform, supplemented by premium access for power users seeking advanced AI models or more tokens. The company also anticipates significant improvements in compute cost efficiency, aiming for a tenfold reduction within the next year, which is essential for achieving profitability at scale. Muse is also slated to play a central role in Meta's smart glasses strategy, indicating its foundational importance to the company's broader hardware and metaverse ambitions, even as the Reality Labs division continues to incur substantial losses.

Key points

  • Meta's AI app, Muse, has reached over 10 million users, with 25% engaging daily.
  • Shopping is a top activity, with engaged users averaging over $200 per purchase, surprising industry insiders.
  • Muse also handles personal productivity tasks like email and calendar, leveraging "Muse Secure VM" for data protection.
  • Meta projects Muse to generate more revenue per user than its core apps and achieve profitability after scaling.
  • The company aims to reduce Muse's compute costs by over 10x in the next year and integrate it into its smart glasses strategy.
The Upside

Muse's rapid user growth and high-value shopping activity suggest a strong path to profitability for Meta's AI ventures. With projected revenue per user surpassing existing apps and significant cost efficiency improvements, Muse could become a major new revenue stream, bolstering Meta's financial performance and validating its substantial AI investments.

The Downside

Despite the positive outlook, Muse's current compute-intensive nature poses a significant challenge, requiring a tenfold cost efficiency improvement to achieve profitability. Furthermore, the Reality Labs division, where Muse will play a central role, continues to incur billions in losses, indicating that Meta's broader AI and metaverse strategy still faces substantial financial hurdles.

Market signals

META· NASDAQ
  • META Meta's AI app Muse has surpassed 10 million users and is projected to generate more revenue per user than its core apps, indicating strong growth potential and a new revenue stream.

AI-generated analysis of potential market relevance. Not financial advice.

Originally reported at

techcrunch.com

Discernion covers the story. Read the full piece at the source.

Tagsaiai-appse-commercemetatechstartups

Author

Sarah Perez

Intelligence analysis by

Gemini 2.5 Flash

Published

Oct 21, 2026

Source

techcrunch.com

Share

Topics

aiai-appse-commercemetatechstartups

Related

More from this desk

Oct 21·techcrunch.com

Recursive launches Flint, a laptop coding model it says was largely built by AI

Recursive, an AI startup, has launched Flint, a free coding model designed to run on laptops and largely built by its own AI research system, the Eureka Machine.

Diagram of nine loops
Oct 21·anthropic.com

Yes Claude Can Do Nine Loops

Physicist Matt von Hippel challenges AI companies to solve a complex physics problem, issuing a challenge to compute scattering amplitudes to nine loops in N=4 super Yang-Mills theory.

A screenshot of two AI-generated images on Just Eat - one of chip muffins - a muffin with potato chips in them, and the other of a pint glass of fizzy drink with an egg and bacon in it
Oct 21·bbc.co.uk

Just Eat apologises for AI food mistakes after social media ridicule

Just Eat apologises for AI-generated food images that led to humorous and sometimes confusing menu items. The company is removing incorrect images and will stop making more until the AI is further developed.

Oct 21·huggingface.co

Profiling TorchNeuron (Part 1): System-level profiling on AWS Trainium

This post teaches how to profile inference performance using AWS Trainium and TorchNeuron, a PyTorch backend for deep learning.