Meta's AI Spending: The $65 Billion Bet on Future Returns

Meta Platforms plans to spend up to $65 billion on artificial intelligence this year, more than double its 2023 capital outlay. The figure, disclosed in the company's latest guidance, covers data centers, GPU clusters, and model training infrastructure. Some investors have started to question the pace, pointing to the lack of a clear revenue line tied directly to the foundational AI work.

The company generates most of its money from advertising. AI features – recommendation engines, automated bidding, content ranking – already feed that business. Meta said these tools boosted ad impression growth by more than 20% in the first quarter. The question is whether that improvement justifies spending at the current rate.

Analysts at several firms have flagged the risk that capex could outpace the revenue uplift for several consecutive quarters. Operating margins narrowed to 38% in Q1 from 42% a year earlier, partly driven by higher depreciation and infrastructure costs. The stock has risen 15% in 2025, trailing the Nasdaq 100's 18% gain.

The alternative view is that Meta is buying an option on the next computing platform. The company's Llama models are open-source, which limits direct monetization but encourages ecosystem adoption. Meta has also invested in AI-powered tools for businesses, including automated ad creation and customer service bots, though these are early in their revenue contribution.

For the thesis to hold, ad revenue growth needs to stay in the mid-teens while capex as a percentage of revenue begins to drop. A management signal that spending will remain elevated into 2026 without a matching revenue acceleration would weaken the case. Conversely, evidence that AI spending is lifting return on invested capital – or the launch of a paid tier for Llama – would support the current valuation.

Meta reports second-quarter earnings on July 30. The number that matters most may not be the top line but the operating margin and the story management tells about when AI dollars turn into AI earnings. At the current price of $568.43, the market is pricing in a continuation of the ad cycle. The alpha score from AlphaScala sits at 56, reflecting a balanced risk-reward. The next earnings call will tilt that either way.