Business · Companies · 2 hrs ago
Meta plans a network of AI agents as spending pressures grow
Meta CEO Mark Zuckerberg wants its Muse AI agents to work together in a large network serving people and businesses worldwide.
He says Meta aims to make Muse available to billions of people and is not charging everyone upfront to encourage use.
Muse launched in early September, and Meta said more than 10 million people in the US and Canada had set up personal agents.
The expansion comes amid privacy and security concerns, including complaints about access to private messages and personal information, as well as a patched security flaw.
Meta also faces a $25 billion New Mexico court judgment over misleading consumers about data privacy in the Cambridge Analytica case, which the company says it will appeal.
The AI push is adding to heavy costs: Meta reported negative free cash flow of $2.12 billion for the quarter ended September 30 and expects to spend at least $135 billion this year, mostly on AI infrastructure.
Meta plans a new class of non-voting stock next year to give it more financing options without reducing existing shareholders’ voting power.
Meta CEO Mark Zuckerberg says the company wants to build a large network of Muse AI agents that work together to serve people and businesses.
Meta says more than 10 million people in the United States and Canada had set up personal agents since Muse debuted in early September.
Zuckerberg said Muse is very compute-intensive, while Meta expects to spend at least $135 billion this year, mostly on AI infrastructure.
Meta reported negative free cash flow of $2.12 billion for the third quarter, which ended September 30.
The company plans to distribute a new class of non-voting stock to existing shareholders starting next year to create more financing options.
- Who
- Meta CEO Mark Zuckerberg and Meta. The company’s Muse AI agents are central to the plan.
- What
- Meta plans to build a large network of AI agents that work together. It also plans a new class of non-voting stock to help finance investments.
- When
- Zuckerberg discussed the plan on an earnings call reported on October 21 and 22, 2026. Meta says the stock distribution will begin next year.
- Where
- The agents are intended to serve people and businesses worldwide.
- Why
- Meta says agents working together could get more done for people and businesses. The stock plan would give the company more ways to raise financing.
Meta
Privacy critics and other critics
Privacy and data handling
Meta
Meta wants to grow Muse and create a network of agents working together.
Privacy critics and other critics
Users have alleged that agents analyzed private texts without permission and revealed a home address; researchers have also raised concerns about detailed personal files.
Use of customer data
Meta
Meta aims to expand Muse, including its shopping use.
Privacy critics and other critics
Amazon blocked the agents from shopping on its platform, citing concerns about handling customer credentials and account data.
Muse debuted, according to the article.
Meta recorded free cash flow of negative $2.12 billion for the quarter.
Meta reported its third-quarter results and discussed its AI plans on an earnings call.
Meta plans to distribute a new class of non-voting stock to existing shareholders.
- Muse agents
- More than 10 million people in the United States and Canada had set up personal agents
- Meta spending outlook
- At least $135 billion this year, with most going to AI infrastructure
- Third-quarter free cash flow
- Negative $2.12 billion
- Planned stock distribution
- Two-for-one distribution of non-voting capital stock to existing shareholders
- Shopping on Muse
- More than a quarter of users engaged in shopping; average purchase size was over $200
3 sources
What people are saying
We expect to grow Muse to serve billions of people. This is both our mission and our strategy
This is also our strategy because we think that billions of Muses in a network at scale will be able to get more done for people and businesses by working together, so we’d like to create a large-scale Muse network as quickly as possible.
Muse is already “very compute intensive to run