Meta's Muse AI Hits 10 Million Users With $200 Shopping Carts

Meta told analysts on its third-quarter earnings call Wednesday that its AI agent Muse has surpassed 10 million users, a figure that confirms independent estimates released earlier in the week by market intelligence firm Sensor Tower.

The headline number is not what caught CEO Mark Zuckerberg's attention. The spending is. Zuckerberg said over a quarter of Muse's engaged users who shopped through the AI agent made average purchases exceeding $200 — a basket size that surprised even industry insiders.

What did Plaid see that no other service shows?

Zuckerberg leaned on a third party to make the spending claim credible. "Plaid told us that they hadn't seen activity like they're seeing with their connector on any other service before," he said, referring to the fintech company that links payment apps to users' bank accounts. Muse uses Plaid's connector to enable agentic shopping — meaning the AI, not the user's thumbs, executes the transaction.

The retention picture is the second concrete signal. Zuckerberg said about 25% of people who set up Muse now use it daily, a proportion he credited with improving retention. For context, a one-in-four daily-active ratio this early in an AI product's life is the metric investors typically demand before believing scale stories. Daily use, however, says nothing about whether those users spend again.

Why does email access matter more than shopping?

Shopping is not the only popular use case. Zuckerberg said users commonly connect Muse to their email and calendars for personal productivity tasks. He framed this as evidence of trust in Muse's security setup, which includes a feature called "Muse Secure VM" designed to protect data and privacy. A forthcoming "Muse Confidential VM," he added, will prevent even Meta from seeing users' content.

That last point carries a caveat worth noting: a promise about a product that does not yet ship is a promise, not a capability. Meta's business model has historically rested on seeing user data to sell ads, which makes the confidentiality pledge a notable — and unverified — departure.

How will Meta make money from Muse?

With Muse having achieved "product-market fit," in Zuckerberg's words, Meta's focus shifts to scaling the agent, growing revenue and cutting compute costs. The company forecasts Muse will eventually generate more revenue per user than its "family of apps" — Facebook, Instagram, WhatsApp and Messenger — which currently brings in "several hundreds of dollars annually" per user, Zuckerberg said.

The monetization structure is specific:

• Muse stays free for most users, with Meta keeping a percentage of commerce transacted on the platform.
• Power users pay for access to the most advanced AI models and more tokens.
• Zuckerberg said Meta believes Muse will be profitable after it scales.

"In the near term, Muse is very compute-intensive to run, but we expect to improve its cost efficiency by more than 10x in the next year, which will enable it to be profitable once the revenue lines are at scale," he told bankers and analysts. He also said Meta is building compute capacity to serve more than a billion people over the next few years.

A 10x cost-efficiency target in twelve months is aggressive by any engineering standard, and Zuckerberg offered no breakdown of how Meta gets there. Profitability, by his own framing, depends on that efficiency claim holding.

What does Muse mean for Reality Labs?

Muse will also anchor Meta's smart glasses strategy. Reality Labs, the division behind smart glasses, VR and the metaverse, is still losing billions — but Zuckerberg said its losses should improve from $19.1 billion in 2025 to $15 billion this year, a narrowing of roughly $4 billion. If agentic AI drives glasses adoption the way Meta projects, Muse becomes the software engine for hardware that has so far burned cash. The scale targets, though, still outrun the revenue lines that would fund them.