Meta is up against a wall of AI pessimism heading into earnings

By Christine Ji

Following Alphabet's earnings last week, investors will be closely scrutinizing Meta's AI spending

Shares of Meta are down 10% on a year-to-date basis.

As Meta Platforms prepares to report second-quarter earnings, the company is confronting a wall of artificial-intelligence pessimism that will be difficult to overcome.

Meta (META) is set to report earnings Wednesday afternoon, along with Microsoft (MSFT), and investors will have an opportunity to judge whether AI momentum outweighs spending fears.

Last week, Alphabet (GOOGL) (GOOG) kicked off the second-quarter earnings season with impressive cloud growth, but spooked the market by raising its capital-expenditure outlook and reporting its first-ever quarter of negative free cash flow.

As a result, Wall Street is recalibrating expectations for the rest of the hyperscalers. Prior to Alphabet's announcement, Deutsche Bank analyst Benjamin Black wrote: "We think there were limited expectations for Meta to increase its current 2026 guidance of $125 to $145 billion," in a note last week. However, investors are now likely bracing for the possibility that Meta could raise its capex outlook.

Meta has been consistently criticized for its spending levels over the past few quarters as the company has overhauled its AI strategy with the creation of Meta Superintelligence Labs. Shares of Meta have dropped 10% since the beginning of the year.

If Alphabet's strong results last week couldn't boost the AI trade, "then even bullish quarters, guidance and AI spending commentary from Meta and Microsoft probably won't appease the doomsayers either," Paul Meeks, managing director at Freedom Capital Markets, wrote in a Monday note.

Analysts tracked by FactSet are anticipating Meta to report revenue of $60.2 billion and earnings per share of $7.19. The company is also expected to see its free cash flow turn negative this quarter, with analysts projecting $801 million in negative free cash flow.

Meta has been directing its spending toward building out more AI revenue streams in addition to its core advertising business. After debuting the Muse Spark 1.0 model in April, Meta followed up with the release of Muse Image and Muse Spark 1.1 in July.

Muse Spark 1.1 was the company's first AI model with competitive agentic coding capabilities and the first to be monetized through an API.

The increased cadence of releases "supports our view that Meta Superintelligence Labs has moved beyond rebuilding the company's training and post-training stack and into a faster, more predictable product cycle," according to Deutsche Bank's Black.

Meta is also reportedly planning to sell compute capacity to third parties, meaning that the company could monetize its data centers beyond its internal workloads.

Such an initiative would be a game-changer for Meta's AI strategy, providing an "off-ramp" in the case that Meta's internal AI models or business agents "fall short of expectations," Raymond James analyst Josh Beck wrote in a note last week. The AI cloud market boasts attractive returns on investment, Beck added, as demand for compute capacity continues to outpace supply.

-Christine Ji

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

07-28-26 0817ET