By the numbers, Meta Platforms had a good quarter. Revenue of $60.8 billion in the three months ended June 30 beat the $60.24 billion analysts had expected, and advertising sales grew across all regions. Investors sold the stock anyway, knocking it down more than 10% in after-hours trading.
The problem was on the other side of the income statement. Earnings of $6.18 a share fell from $7.14 a year earlier, and free cash flow dropped sharply as Meta poured money into the data centers, chips and research that its artificial-intelligence push requires. Management offered no timeline for when that spending would produce returns on the scale investors want.
The third-quarter outlook added to the unease. Meta guided to revenue of $61.5 billion to $64.1 billion, with a midpoint of about $62.5 billion, below the $63.17 billion analysts had projected. The shortfall, small in dollar terms, was read as a sign that the AI boom has not yet translated into a commensurate bump in ad prices or engagement.
Wall Street’s reaction was swift. Analysts quoted in coverage of the results said Meta’s AI investment has moved beyond what the company can comfortably absorb, and Morgan Stanley downgraded the stock, arguing that capital spending is growing more than twice as fast as revenue. The spending plans, Morgan Stanley said, have shifted the risk in Meta’s story from execution to affordability.
Meta’s position is unusual among the big AI spenders. Unlike Microsoft, Alphabet and Amazon, whose AI investments feed cloud businesses that charge customers directly, Meta’s AI spending is largely a bet on improving its own products: sharper ad targeting, better content recommendations and a family of consumer AI assistants that so far generate little direct revenue.
The company argues the calculus is simple. AI-driven recommendation improvements have already lifted engagement, and advertisers are paying more for campaigns that use its AI tools. Executives have said the spending is a down payment on a future in which AI touches every part of the business, and that competitors doing less will fall behind.
Investors have heard that argument before, and the market’s reaction on Thursday suggests patience is thinning. Meta’s shares had risen more than 60% over the past year on the strength of its advertising machine and its early lead in consumer AI, and the multiple left little room for a quarter that failed to confirm the narrative.
The comparison with peers cuts both ways. Meta is spending heavily on AI with none of the cloud revenue that justifies similar outlays at Microsoft and Alphabet, and its capital expenditure is now several times what it was two years ago. But Meta’s core business remains exceptionally profitable, and its ad machine has repeatedly surprised to the upside.
The free cash flow decline is the number to watch. Meta burned through its cash cushion as it pays for data centers under construction, and a year of spending at this pace will test even its balance sheet. The company has said it can fund the buildout from operations, but the arithmetic leaves less room for the buybacks that have supported the stock.
Meta’s history offers a template for the current moment. The last time the company’s spending spooked Wall Street, in early 2022, the shares fell more than 60% before a grueling cost-cutting campaign and a rebound in advertising restored investor confidence. Executives have said they learned that lesson and will run the AI buildout with tighter discipline, but Thursday’s drop showed how quickly the market reverts to the old playbook.
Analysts who remain bullish argue the selloff is a buying opportunity. They note that Meta’s AI investments are showing up in engagement data, that the company’s family of apps reaches billions of users, and that ad spending follows attention. The bear case is equally direct: if AI does not deliver a step change in revenue growth within a few quarters, Meta will face the choice of cutting spending or disappointing investors on margins.
Meta’s spending plans are set for the year. The company has said capital expenditures for 2026 would rise well above last year’s level as it builds the data centers and buys the chips its AI ambitions require, and the second-quarter cash flow figures show the bill arriving. The Reality Labs division, home of the company’s mixed-reality hardware, continued to post large operating losses, adding a second drain alongside the AI buildout.
The product side offers some offset. Meta has pushed AI features into its family of apps, including assistants inside WhatsApp and tools that summarize information across its platforms. Engagement data improved in the quarter, executives said, and advertisers using the company’s AI campaign tools are seeing better returns, which is the mechanism by which the spending eventually pays for itself.
Thursday’s decline was the market voting on which story it believes. Meta has weathered skepticism before and emerged with earnings growth intact. Whether this cycle repeats depends on a question the company has not yet answered: when the AI spending starts paying for itself.
