Meta just posted one of its strongest advertising quarters in years, and the numbers reveal something bigger than one company’s performance: the ad platforms marketers depend on are increasingly using record ad prices to fund an AI infrastructure buildout that has little to do with serving campaigns better.
The quarter, in the numbers that matter
Meta’s total revenue reached $60.80 billion in the second quarter of 2026, up 28% year over year, according to the company’s earnings release filed with the Securities and Exchange Commission. Advertising revenue, the piece marketers actually pay into, grew 27% to $59.36 billion. That growth came from two levers: ad impressions rose 14% and the average price per ad climbed 12%. Family daily active people reached 3.60 billion, up a modest 3%, meaning most of the revenue gain came from charging more per ad rather than reaching more people. “AI is accelerating our core business today, powering our next generation of products, and opening the door to entirely new enterprise opportunities,” CEO Mark Zuckerberg said in the release.
Where the extra dollars are going
Total costs and expenses jumped 55% to $42.03 billion, a pace that far outstrips revenue growth. Roughly $2.40 billion of that was legal charges and $1.18 billion was severance tied to a May 2026 headcount reduction affecting about 8,000 employees. Capital expenditures for the quarter hit $31.08 billion, and Meta raised its full-year capex guidance to a range of $130 billion to $145 billion, narrowed upward from its prior $125 billion to $145 billion outlook. Full-year total expenses are now guided to $165 billion to $169 billion.
The margin story underneath the headline growth
The clearest evidence of what is happening sits below the revenue line. Operating margin fell to 31%, down from 43% in the same quarter a year earlier, and operating income actually declined 8% even as revenue grew 28%. Net income dropped 14% to $15.85 billion, and diluted earnings per share of $6.18 missed the pace that top-line growth alone would suggest. Cash flow from operations was a healthy $31.86 billion, but capital expenditures of $31.08 billion consumed nearly all of it, leaving free cash flow of just $784 million, a fraction of what a company growing revenue this fast would typically generate. That gap between operating cash and free cash flow is the AI buildout showing up on the balance sheet in real time.
Why investors are not celebrating
Meta’s shares fell roughly 8% to 10% in after-hours trading following the July 29 report, a reaction that has less to do with the ad numbers than with what backs them. Microsoft’s cloud business grew 43% year over year and Alphabet has Google Cloud, giving both companies a visible, separately reported revenue line that scales alongside their AI spending. Meta’s AI investment instead shows up folded into the existing ad engine: real, and measurably stronger, but with no distinct line item investors can point to as proof the spending pays for itself beyond juicing ad prices. Zuckerberg’s stated thesis, that selling AI-driven intelligence eventually carries better margins than the compute costs behind it, remains a thesis rather than a reported number. For marketing leaders, that is the tell: the AI premium showing up in CPMs today is financing a bet the market itself has not yet agreed to underwrite.
What this means for the marketing leader
The mechanism here matters more than any single quarter. When a platform needs to fund tens of billions in AI infrastructure, and impressions cannot grow indefinitely because attention is finite, price per ad becomes the lever platforms pull. A 12% price increase on the same inventory is effectively a tax the ad ecosystem pays toward compute capacity, not toward better targeting or better creative. Marketing leaders who budget for flat CPMs next planning cycle are working from an outdated assumption.
This is not isolated to Meta. Amazon’s own Q2 advertising business crossed nearly $20 billion in the same quarter, evidence that the pattern of platforms leaning on ad pricing to fund AI ambitions is systemic across the major buying channels, not specific to one company’s roadmap.
The efficiency question gets harder, not easier
As price per ad rises, the pressure to prove efficiency rises with it. Automated bidding systems already make most of the moment-to-moment buying decisions on these platforms, and AI-driven bidding has become table stakes across search advertising, which means marketers increasingly cannot see, let alone negotiate, the mechanics behind a price increase. The optimization layer and the price-setting layer now sit inside the same black box.
That combination, rising prices plus reduced visibility into why, changes what marketing leaders should be asking vendors for. Reach and impressions are no longer the right questions on their own. The better ones are: how much of this quarter’s CPM increase reflects genuine demand versus platform-side cost recovery, and what guardrails exist to cap spend if automated bidding chases price rather than outcomes.
What to do about it
Three moves make sense heading into next quarter’s planning. First, build AI-driven capex cycles into forecast models the same way seasonality is already built in; a platform whose free cash flow nearly vanished under $31 billion in quarterly capex, as Meta’s just did, is not going to absorb that cost itself. Second, push media partners for outcome-based reporting rather than reach metrics, since impression growth is now the smaller half of the revenue story. Third, treat channel diversification as a hedge against pricing power concentrated in a small number of platforms that are all funding similar infrastructure bets at the same time.
The AI story on these platforms has mostly been told as a product narrative: new agents, new creative tools, new targeting. The Q2 numbers tell the financing side of that story, and marketers are the ones underwriting it.
Source: Meta Reports Second Quarter 2026 Results (SEC filing)
