Muse Is Meta’s Biggest Bet Yet
Meta positions Muse as a personal AI agent that can shop, work, and communicate with other people’s Muses on our behalf. More than 10 million users in the United States and Canada have already downloaded and set it up.
But user numbers are not the only deciding factor. Privacy remains a concern, with reports of agents leaking information from private conversations and an incident in which a transaction sent a stranger to a user’s home.
Meta must also prove that Muse is worth the investment. The company plans to spend between $135 billion and $145 billion on AI data centers in 2026, while user and business trust cannot be forced so easily.
Where Is Muse Taking Meta?
Meta positions Muse as a personal AI agent and the starting point for a “large-scale Muse network,” where each person’s agent can communicate and conduct transactions with others. In Zuckerberg’s vision of the future, Muse will help people save time, save money, build businesses, and shop on their behalf.
The early signals are fairly clear: More than 10 million people in the United States and Canada have downloaded and set up Muse. One in four users shops through the platform, spending an average of more than $200 per transaction.
The Day AI Agents Start Talking for Us
In one case, Muse handled a transaction on Facebook Marketplace on behalf of a user but ended up sending a stranger to the user’s home in the middle of the night.
The incident raises an important question: When AI has permission to contact other people and handle matters for us, could convenience also bring risks that humans never anticipated?
From Meta’s Money-Making Apps to the Muse Network
Today, Instagram and Facebook remain Meta’s primary revenue-generating platforms. But Zuckerberg envisions Muse as something much bigger: a personal assistant that can communicate with other people’s Muses and handle transactions on users’ behalf.
Meta’s vision is that as more people use Muse to work, shop, or create new things, revenue per user could surpass that of Instagram and Facebook. Muse could then become a highly profitable business within a few years—making it a major bet on the company’s future.
How Is Muse Different from What Meta Already Has?
This is a comparison between Meta’s “existing platforms” and its “new bet,” not a comparison of product versions.
| Factor | Instagram and Facebook | Muse |
|---|---|---|
| Revenue source | Generate nearly all of Meta’s current revenue | Collect transaction fees from commerce |
| Role for users | Meta’s primary platforms | personal AI agent that helps with work, shopping, and communicating with other Muses |
| Business direction | Existing businesses that already generate revenue | A bet on the Muse network and future commerce |
When Muse Starts Working in Everyday Life
If users need to save time, Muse could help manage personal goals, recommend options, or handle certain tasks on their behalf, operating 24/7.
When users want to buy something, Muse does more than take them to a website. It can research information, recommend products, and communicate with a store’s agent to complete the transaction on the user’s behalf.
From a business perspective, a buyer’s agent will communicate directly with a store’s agent, allowing product selection and transactions to take place through Muse.
If multiple users have their own Muses, these agents could coordinate purchases or communication between people without requiring the owners to handle every step themselves.
Muse’s Competitors That This Article Cannot Yet Answer
The sources do not identify any competitors or alternative products, so it is not yet possible to conclude which products Muse outperforms. The comparison available for now is between Meta’s promises and the evidence currently available.
| Factor | What Muse promises | Evidence available now |
|---|---|---|
| Personal use | Help save time and money, and create new things | Identified daily users, but no retention details |
| Shopping | Agents communicate with stores and complete transactions on users’ behalf | Users shop through the platform, but the actual workflow remains unclear |
| Privacy | Built for privacy and security | Reports of agents forwarding private content and incidents involving errors |
Muse has promising early signals: More than 10 million people in the United States and Canada have downloaded and set it up, and one in four downloaders uses it daily. Meanwhile, one in four users shops through the platform, with average spending of more than $200 per transaction.
Pros
- +Has daily users and real shopping activity
- +Supports transactions averaging more than $200
Cons
- −Reports of agents leaking content from private conversations
- −Still exposed to financial-data risks and uncertain business adoption
The Price That Does Not Appear on the Screen
Muse may be free for consumers, but businesses will be responsible for transaction fees. Meta will collect fees from purchases made on the platform, so costs could shift from storefronts into the AI-agent system instead.
The risks are not limited to cost. Agents can search for, recommend, and order products at any time, creating opportunities for mistakes or problems for users—similar to the reported case in which Muse sent a stranger to someone’s home at night while handling a Facebook Marketplace transaction.
Privacy is also a concern. There have been reports of agents leaking content from private conversations, and Canadian banks may block access to customers’ financial data.
Behind the scenes, Meta plans to spend $135–145 billion in 2026—twice as much as the previous year—along with a $25 billion obligation to New Mexico.
If this future becomes reality, who will need to adapt first?
Ultimately, will Muse’s future be determined by AI capabilities or by trust from users and businesses? Allowing an agent to shop and manage information on someone’s behalf requires more confidence than promises alone can provide.
User numbers, real-world usage, and risk management will determine whether Meta can take Muse to the AI network it has envisioned.