Meta's advertising business is on a tear, fueled by better monetization of Reels, a boom in click-to-message ads, and AI-powered tools that sharpen ad targeting. But the company is pouring billions into artificial intelligence infrastructure — data centers, chips, and model development — and investors are starting to ask whether the ad growth can keep up with the spending.
What's driving the ad growth
Revenue from Reels is closing the gap with Feed and Stories, though it hasn't fully caught up yet. Meta says further improvements are expected. Click-to-message ads, which let users interact with businesses on WhatsApp, Messenger, and Instagram Direct, have become a multi-billion-dollar run-rate business. The company sees potential for AI to automate more of those customer interactions, which could drive even higher returns.
Meta's Advantage+ suite of AI tools has helped recover ad performance after Apple's App Tracking Transparency changes. The tools use dynamic creative, broad targeting, and model-driven budget allocation to deliver better results. That's been a key factor in keeping advertisers on the platform.
The cost of the AI bet
Meta is significantly increasing capital expenditure for AI infrastructure. That includes building training clusters, inference capacity, and a software stack to improve recommendations and ad outcomes. The company has a multi-year build plan, and the financial impact is already visible: upfront capex, depreciation, higher operating expenses from hiring and model operations, and pressure on cost of goods sold from running inference at scale.
The release of Llama 3, Meta's open(ish) large language model, signals a commitment to improving its own ranking, safety, and ad products through ecosystem testing. But the spending is a bet that these investments will eventually pay off in higher ad yield and new revenue lines — enterprise tools, API usage, and subscription value-adds.
Investor concerns
The tension in Meta's earnings is between strong ad growth and rising AI spending. Investors are watching whether the ad business can sustain the capex cycle without squeezing margins, buybacks, and profitability. Meta's revenue is overwhelmingly from ads across Facebook and Instagram; new formats are less important than the underlying pricing and delivery systems. The question is whether AI spending will boost those systems enough to justify the cost.
Meta's next earnings report will be the first real test of whether the AI investment narrative holds up against the numbers. The company has to show that the billions going into data centers and model training are translating into better ad performance and new revenue streams — not just higher costs.
