The company also said it would issue a new class of stock to raise more cash for its AI efforts.

Investors ignored the short term pain for the potential of a future AI payoff, cheering early signs of traction with the recently released Muse AI agent, sending Meta shares up as much as 5% in after hours trading on Wednesday.

“Our investments are paying off, the core business is strong, and we’re seeing velocity across the board,” Zuckerberg said in a prepared statement on Wednesday, later adding on the conference call that plans were underway to monetize Muse.

“The average revenue per American user for our existing Family of Apps users is several hundreds of dollars annually, and we expect Muse will reach an even higher level over time due to the amount of commerce that we already see taking place,” Zuckerberg said.

Meta’s third-quarter revenue rose 26% to $64.6 billion, above the $63.3 billion analysts expected, according to Visible Alpha consensus from S&P Global Market Intelligence. And the company forecast revenue for the final quarter of the year to range between $73 billion and $76 billion, compared to the $73.75 billion expected by analysts.

But Meta’s operating income of $13 billion was down 37% from the $20.5 billion it posted in the second quarter of 2025 as a result of a $10 billion settlement Meta struck with a group of attorneys general over youth safety. The terms of the settlement will see Meta pay about $18 billion to the states over the next decade, depending on cooperation from TikTok and Google’s YouTube. Without the charge, operating income would have been $22.9 billion and earnings of $7.40 a share, rather than $13 billion and $3.83.

Legal issues may continue to take a bite out of Meta’s results. Hours before Meta announced its Q3 results Wednesday, news broke that a New Mexico judge ordered the company to pay $25 billion in penalties related to the Cambridge Analytica data privacy scandal. Meta plans to appeal.

Net income in Q3 totaled $9.9 billion, or $3.83 a share, versus $2.7 billion, or $1.05 a share in the year-ago period.

Meta is plowing unprecedented amounts of capital into AI infrastructure and services, joining companies like Alphabet, OpenAI, and Microsoft in a race to develop the most advanced AI models. To continue funding the AI effort, Meta said it would re-allocate investments from its money-losing virtual reality business, Reality Labs, and that it would create a new non-voting class C common stock.

Capital expenditures, which have tech investors on tenterhooks, clocked in at $34.9 billion, compared to $31 billion in the second quarter and $19.4 billion a year ago. Meta slightly tweaked its full year capex guidance to $135 billion to $145 billion, from $130 billion to $145 billion. 

Free cash flow, which has been a sore spot for Meta, was negative $2.1 billion, compared to $784 million in the second quarter and $10.6 billion a year ago. 

Monetizing the Muse

The results also come a week after Meta’s Superintelligence Labs launched Muse Wonder to the public, its first frontier model, and a little over a month since Meta released its Muse virtual assistant.

Zuckerberg said that more than 10 million people in the U.S. and Canada had already set up Muse assistant since it was released. “Because this is a compute intensive service to run, we have not launched it globally yet,” Zuckerberg said on the earnings conference call with analysts.

Asked about Meta’s traditional playbook of monetizing new products after they reach 1 billion users, Zuckerberg said that the steep costs of running Muse and the inherent suitability of AI agents like Muse to play a role in retail transactions meant that the company would look for ways to generate revenue sooner.

“My guess is, I think that we are going to work on monetization somewhat sooner on this than our normal playbook,” he said.

Zuckerberg also said he was optimistic about the potential for business customers to use Muse.

Few companies are as as well positioned as Meta to capitalize on the agentic AI opportunity, Zuckerberg said, though he noted that the company needed to continue making progress on compute efficiency.

“It’s going to be very important to get the math on all of this to work out,” he said.

Melissa Otto, global head of Visible Alpha research at S&P Global, wrote in a note before the results that consensus capex for 2026 had risen more than $14 billion since April to about $140 billion—almost double 2025. She noted that investors continue to want clarity on how that spending will translate into shareholder returns. 

Overall in tech infrastructure spending, capex is expected to continue ballooning in 2026 to $790 billion, according to Otto, almost double what it was in 2025 at $404 billion. And it won’t stop there. For 2027, capex is expected to rise to $1.2 trillion, meaning companies will have spent nearly $2 trillion on the AI buildout in two years.