Meta Stock After Q3 2026 Earnings: Price Per Ad Now Grows Faster Than Impressions

Meta reported its third quarter after the close on October 21. Revenue grew 26% to $64.6 billion, ahead of the roughly $63.4 billion analysts expected and above the top of Meta's own $61 to $64 billion range. GAAP earnings per share were $3.83 because of a $9.89 billion charge for the August settlement with state attorneys general; without it, EPS was $7.40. Meta guided Q4 revenue to $73 to $76 billion. The stock was up about 3% after hours.

In July I wrote that Meta's AI build has nobody to sell capacity to, so the only place it can show a return is the price of Meta's own ads. Impressions were slowing because Meta's audience only grows about 3% a year. If the models were working, price per ad would have to pick up the slack. I said the thesis would break if price per ad grew less than 8% for two straight quarters while impression growth fell into single digits.

A year ago impressions were doing most of the work. In Q4 2025 they grew 18% and price per ad only 6%. Since then impression growth has fallen by almost half, and price growth has more than doubled. Ad revenue grew 26% in Q3, close to Q2's 27%, so the slowdown in volume didn't reach the top line. The price side covered it.

This was also the quarter the bill got bigger. This was the first quarter of the build with negative free cash flow, and Meta is now borrowing to pay for it. I don't think that changes the July argument, which was about whether the spending shows up anywhere. Whether it pays back is a later question. But it does mean the price line has to keep going for years, not quarters.

Here's the problem with calling this a win. If advertisers keep spending roughly what they planned and Meta sells fewer new ad slots, the auction clears higher. Price per ad goes up without the ads getting any better. That's what you'd expect when impression growth halves, and it's exactly what the chart shows. It also happened before: price per ad grew faster than impressions through 2024 and into Q1 2025, when impression growth was in single digits. I raised this in July and I can't rule it out with one more quarter of the same shape.

There's one reason to think it's more than scarcity this time. Meta's revenue beat its own guidance, which means advertiser budgets grew faster than Meta expected in July. Scarcity alone doesn't explain bigger budgets, but better results per ad dollar would. That's my reading, and I can't prove it. Meta doesn't publish conversion rates, and the bigger budgets could also reflect a strong ad market across the board.

The cleaner tests sit outside Meta. AppLovin owns no inventory and lives on matching, and it reports next month. If its net revenue per install keeps rising, that's evidence matching is what advertisers pay for. The Trade Desk is the control. It sees the bid but not the sale, and it guided Q3 to at least $650 million, below the roughly $739 million of a year earlier. If it comes in well above that guide, ad budgets are strong across the open web too, and Meta's beat tells me less about matching than I'd like.

July named this case but didn't put a number on it. Here's one. If Meta finds more inventory in 2027, through Threads, WhatsApp, or ads in Muse if it adds them, and impression growth climbs back above 12% for two quarters, a matching gain should keep price per ad growing. If price growth instead drops below 8% as soon as supply loosens, the 2026 price gains were the auction clearing higher, and I'd have read them wrong. If impressions never get back to 12%, the July test still applies: price per ad under 8% for two quarters while impressions stay in single digits. I'll update this when Meta reports Q4 in late January, and when 2027 capex is finally guided.