How I Made $5000 in the Stock Market

Meta Stock Tumbles After Earnings. Why AI Costs Are Becoming a Problem.

Oct 29, 2025 02:30:00 -0400 by Adam Levine | #AI #Earnings Report

Mark Zuckerberg, CEO of Meta Platforms, wearing the company’s Orion augmented reality glasses at a company event in September. (David Paul Morris/Bloomberg)

Key Points

Meta Platforms reported mixed third-quarter earnings results Wednesday afternoon. Its shares were down over 11% by midmorning Thursday.

Earnings-per-share were $1.05, well behind Wall Street’s consensus estimate of $6.72, after a large, one-time charge in the quarter. Revenue for the quarter reached $51.2 billion, ahead of expectations for $49.5 billion, and up 26% on the year.

Analysts closely watch Meta’s operating margin, which came in at 40% for the third quarter, down from 43% last year. Meta’s cost growth is outpacing sales, and the company upped its expense guidance for 2025. Research and development expenses led the way, up 28% from last year due to a hiring binge in Meta’s artificial-intelligence research unit.

Created with Highcharts 9.0.1Meta PlatformsStock ticker: METASource: FactSetAs of Oct. 30

Created with Highcharts 9.0.1Aug. 2025Oct.700710720730740750760770780790$800

Meta stock was down as much as 8.5% in premarket trading.

Beyond operating expenses, after raising 2025 guidance for spending on artificial-intelligence data centers twice this year, Meta did it again. It now expects around $71 billion for the year, up from $69 billion. The company also said that “our current expectation is that capital expenditures dollar growth will be notably larger in 2026 than 2025.”

Meta isn’t slowing down.

Despite disappointment on the expense side, Meta’s revenue outlook for the fourth quarter was solid.

Meta’s AI investments have reshaped its balance sheet and cash flow statement. In the second quarter, Meta had more debt, tax, and lease liabilities than cash and short-term investments for the first time, and that deepened in the third quarter.

Even though third-quarter operational cash flows rose by 21% from 2024, free cash flow, after subtracting capex, declined by 35%%.

Last week Meta also arranged a $27 billion financing deal that places its giant Hyperion data center into a joint venture with alternative investment manager, Blue Owl Capital. This will keep the associated debt off Meta’s balance sheet and the capex off its cash flow statement.

Unlike the other big AI spenders, all of Meta’s data center investments are for itself, not for renting out in the cloud as with Amazon Web Services, Microsoft Azure, and Google Cloud Services.

Besides shifting the company in a more asset-heavy direction, CEO Mark Zuckerberg is also upending Meta’s AI research and development. He aims to offer AI to the 3.5 billion average daily users who use at least one of Meta’s apps. The company wants to provide new user experiences like the Meta AI chatbot, while increasing the effectiveness of ad targeting.

“It’s pretty early, but I think we’re seeing the returns in the core business,” Zuckerberg said on the earnings call. “That’s giving us a lot of confidence that we should should be investing a lot more, and we want to make sure that we’re not underinvesting.”

But the development of Meta’s AI language models that underpin this strategy has been halting. The company’s Llama 4 models were released in April to poor reviews, leading Zuckerberg to rethink his AI teams. In June, he began offering the best AI researchers in the world large multiyear compensation packages, reportedly stretching to nine-figures. At the center of the effort was a deal to acquire 49% of start-up Scale AI at a $29 billion valuation; the primary goal of that deal was to hire its CEO, Alexandr Wang, to run Meta’s AI research.

Write to Adam Levine at adam.levine@barrons.com