Updated Version 4 - Meta Puts Muse Proxy at the Core of Its AI Strategy and Plans to Issue New Shares to Drive Investment

Add charts and details

Jaspreet Singh/Katie Paul

Reuters, October 21 - Meta META.O on Wednesday launched a new class of non-voting shares to help fund the social media giant’s costly AI expansion plans, as executives outlined a vision for building the company’s future around its emerging AI agent business.

The parent company of Facebook and Instagram saw its shares rise more than 4% in after-hours trading. Meta also forecasted that fourth-quarter revenue would surpass Wall Street estimates and reported strong growth in its newly launched consumer AI agent app, Muse (link).

The company has been investing heavily in advanced chips and data centers to more aggressively compete with AI market leaders OpenAI and Anthropic, propelling it to the forefront of the emerging consumer AI agent track but also eating into its traditionally strong free cash flow.

Meta will introduce a new class of non-voting Class C common stock in January 2027, effectively constituting a 2-for-1 stock split, aimed at funding acquisitions, retaining top talent, and maintaining CEO Mark Zuckerberg’s control.

Zuckerberg stated that in just a month and a half, over 10 million users in the US and Canada have registered Muse accounts, and the growth momentum remains strong.

He expects Muse will eventually serve “billions of people,” with revenues per user potentially outpacing Meta’s average, predicting the project will generate network effects similar to those that once helped him build his social media empire.

“One reason we’re not charging everyone directly for usage is because we want to build this Muse network, and we believe when a large number of users and businesses engage with Muse, its value will increase with scale,” Zuckerberg said during Meta’s post-Q3 earnings call.

https://www.reuters.com/graphics/META-STOCK/gdpzgadqrvw/chart.png

Higher Costs

Nevertheless, Zuckerberg admitted that launching Muse is more expensive than many other new products Meta has rolled out over the years, saying the company will start commercializing the virtual assistant “a little earlier than usual.”

Meta CFO Susan Li projected that the company’s AI business would start to contribute incremental revenue in 2027 and “become a significant source of revenue diversification over time.”

Executives also emphasized AI as the future development focus for Meta’s loss-making Reality Labs division. This unit once led the company’s push to build the “metaverse”—the immersive digital world that inspired the company’s 2021 rebrand from Facebook to Meta.

In earnings, they stated they “remain committed to improving Reality Labs’ profitability,” forecasting that “with the shifting of investments to other priority areas, primarily AI,” the unit’s losses will narrow to around $15 billion next year.

The company raised the lower end of its annual spending outlook, now expecting this year’s capex to be between $135 billion and $145 billion, up from the previous $130 billion to $145 billion range.

The company reported negative Q3 free cash flow of $2.12 billion, compared to positive $10.63 billion a year ago.

https://www.reuters.com/graphics/META-CASH/zgpodlkrbvd/chart.png

According to data compiled by LSEG, Q3 revenue reached $64.6 billion, beating expectations and exceeding Wall Street’s estimate of $63.37 billion. Meta also projected Q4 revenue to reach between $73 billion and $76 billion, above the $73.73 billion estimate.

The initial popularity of Muse, along with the release of competitive next-gen AI models (link), has helped Meta justify its spending to investors, even as some companies restrict access to the AI agent over safety concerns.

“The market has shifted from punishing Meta for its spending to rewarding it for a compelling product story,” said Jesse Cohen, senior analyst at Investing.com.

Legal Challenges

Legal challenges are mounting, including earlier Wednesday when a New Mexico judge (link) ordered a $25 billion fine—after a jury found Meta misled state residents over how it shared user data with third parties and enforced policies targeting hate speech and misinformation, in a case related to UK political consultancy Cambridge Analytica.

Minda Smiley, senior analyst at eMarketer, said the $25 billion fine was “a stark reminder of the legal challenges Meta currently faces, which will persist into next year.” “If AI agents become more mainstream, privacy and security issues and concerns will only intensify, and given Meta’s reputation in this area, this could ultimately be damaging.”

In August, Meta agreed to pay up to $18 billion over ten years (link) and to restrict teens’ use of Facebook and Instagram, settling allegations from nearly every US state that Meta’s platforms were designed to be addictive to children.

Meta reported diluted earnings per share of $3.83 this quarter—it would have reached $7.40 if not for $9.89 billion in charges from the August settlement.

The settlement removed a major regulatory overhang (link) that had pressured Meta’s stock price and enabled the company to avoid a potentially lengthy trial—which could have exposed more internal documents about its handling of young users.

However, in its Wednesday guidance, Meta cautioned it still faces scrutiny and US litigation around “youth-related issues” that could impact its financial results.

https://www.reuters.com/graphics/META-DAP/bypranklkve/chart.png

(For the convenience of non-native English speakers, Reuters automated these reports into several other languages. As automated translation may contain errors or lack context, Reuters does not guarantee the accuracy of these texts, which are provided solely for reader convenience. Reuters accepts no liability for any damages or losses arising from the use of the automated translation function.)