Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Meta says Muse tops 10 million users, with shopping a leading activity

Meta said on its third-quarter earnings call that its AI app Muse has more than 10 million users, with about 25% of people who set it up using it daily. Shopping through the agent has surprised even insiders, with over a quarter of engaged shoppers averaging purchases above $200.

Published
Meta 称 Muse 用户已超 1000 万,购物成为主要用途
Image source: techcrunch.com

Meta said during its third-quarter earnings call on Wednesday that its AI app Muse has been adopted by more than 10 million users, confirming estimates published earlier in the week by market intelligence firm Sensor Tower.

How people use it matters more than the headline number. CEO Mark Zuckerberg said about 25% of people who set up Muse now use it daily, which has helped retention, and that more than a quarter of engaged users who shop through the agent had an average purchase size above $200.

The shopping flow runs on payments infrastructure. Muse uses Plaid's connector to link bank accounts, and Zuckerberg quoted the fintech company saying it had not seen activity like this on any other service. Connecting the agent to email or calendars for personal productivity is another popular use.

Zuckerberg attributed that willingness to Meta's security design, including a feature called Muse Secure VM that isolates user data, and said a forthcoming Muse Confidential VM will keep content invisible even to Meta.

With Muse now at what Zuckerberg called product-market fit, Meta's focus shifts to scaling the agent, growing revenue and optimizing compute costs. He said Meta expects Muse to eventually generate more revenue per user than its family of apps, which currently brings in several hundred dollars annually per user.

The monetization outline is a free tier for most users, a percentage of commerce conducted through the platform, and paid access for people who want the most advanced models or more tokens. Zuckerberg said Muse should be profitable once it scales.

Compute is the binding constraint. Zuckerberg said Muse is very compute-intensive to run today, but Meta expects to improve cost efficiency by more than 10x in the next year and is building capacity to serve more than a billion people over the next few years.

Muse also sits inside a longer product story: it will play a central role in Meta's smart glasses strategy, while Reality Labs losses are expected to narrow from $19.1 billion in 2025 to $15 billion this year. Zuckerberg teased a Muse network that would let users' agents work with those of friends and family, and let shoppers talk to a business's agent rather than navigate its website.

Why it matters

Muse's numbers turn Meta into a consumer agent platform with a plausible commerce take rate, and put shopping rather than chat at the center of how it plans to make money from AI.

MetaAI Agent电商
Back to realtime news

Nearby Updates

All

10/22, 04:57

DeepSeek open-sources DeepInfer, the engine behind its production serving

DeepSeek has open-sourced DeepInfer, the inference framework that powers its large-scale production services, releasing an offline inference build that currently supports DeepSeek-V4.1-Flash. The code is published under Apache 2.0, with the repository listing missing pieces such as a KV cache system, disaggregated serving and cross-replica load balancing.

10/22, 04:27

Recursive ships Flint, a free coding model built largely by its own AI

Recursive released Flint, a free coding model that runs on a laptop and, according to the company, was built largely by its automated research system, the Eureka Machine. Flint is a 40-billion-parameter mixture-of-experts model with 2.5 billion active parameters, published under Apache 2.0 on Hugging Face.

10/22, 03:53

Anthropic launches Claude Duet, an always-on agent that comes to you first

Anthropic opened a research preview of Claude Duet, an app built around Claude Opus 5.5 in which Claude acts as a proactive agent that monitors connected apps, works on a cloud virtual machine and checks in before sending messages, making payments or deleting anything. It is included in Pro and Max plans for US subscribers via a waitlist, and is not yet available on Team or Enterprise plans.

10/22, 03:50

GPT-5.6 Sol and Gemini 4 Argon found a critical PHP flaw affecting millions of sites

OpenAI's GPT-5.6 Sol discovered a critical PHP vulnerability through the Patch-the-Planet initiative, and security firm Wiz independently found the same flaws using Wiz Atlas with Gemini 4 Argon and GPT-5.6 Sol. The fix took two months of coordinated work across five phases, and OpenAI warns the same capabilities can help attackers.