Meta's Free Cash Flow Dropped 91% on AI Spending. This Week Is the Test.

Meta's cash flow fell sharply, but the core ad engine is still running

In the most recent quarter, free cash flow plunged 91% to just $784 million. That was a sharp drop from the roughly $12 billion per quarter Meta had averaged over the prior eight quarters. The reason is straightforward: the company poured much more cash into servers, data centers, and network infrastructure to fund its AI buildout.

The business itself is still growing. Revenue rose 28%, yet short-term profitability came under pressure as spending climbed. At the same time, 2026 capital expenditure now sits at $130 billion to $145 billion. In simple terms, Meta is generating more ad revenue, but almost all of it is being redirected into AI infrastructure rather than returned to shareholders.

The real test is whether AI improves ad profitability, not just platform scale

Meta's AI spending only works if it makes the ad machine more profitable, not just larger.

Better ad matching should show up in margins

There is at least some early evidence that AI is helping. Meta said AI is improving content recommendations and ad matching, and advertising revenue climbed 27%. The company also produced $31.86 billion in operating cash flow in the quarter, even as it spent about $31.08 billion on capex. That means the core business is still funding almost the entire buildout from current operations. Meta is under investment pressure, but it is not yet in financial distress.

Why investors are split on the spending

This is where the debate sits. If AI acts as a profit multiplier, each extra dollar of investment should eventually generate more than a dollar of better ad monetization. If it behaves more like a utility-style cost base, investors will keep financing a massive infrastructure bill just to keep the platform competitive, with returns coming slowly.

Meta's latest numbers fall into that gray zone. The cash engine and ad demand still look healthy, but the immediate payoff still looks small relative to the scale of spending. Management has already pointed to a capex path of up to $145 billion this year, so the real question is whether AI keeps improving ad performance fast enough to justify that pace.

What investors should watch in this earnings update

Expectations were already tight before results. The market was looking for about $60.23 billion in revenue, with ad growth supported by ad impressions increased by 14% year-over-year and average price per ad increased by 12% year-over-year. If management can show that AI is strengthening both ad volume and pricing, investors are more likely to view the spending as productive. If guidance disappoints, concerns about a slow AI payoff will likely intensify, especially after Meta gave a disappointing revenue forecast for the current quarter.

There is also a possible offset. Meta is in talks to lease computing power to Anthropic in a potential $10 billion deal. If that translates into a real business, some of the infrastructure spend could eventually support a separate revenue stream rather than sitting purely as a cost center.

The clearest signals for or against the thesis

What would support the bulls: - Meta meets or beats Family of Apps operating income estimate - It sustains the impressions and price-per-ad growth pattern - It gives a credible update on capex pacing against the $130 billion to $145 billion outlook - It provides a concrete comment on compute leasing talks

What would strengthen the bears: - A wider miss in Family of Apps profitability - Weaker ad volume or pricing than investors expect - No meaningful update on when the AI buildout will start showing clearer returns - Little progress on turning infrastructure into a revenue opportunity, including the Anthropic compute deal