Newsletter digest — August 4, 2026: Meta's earnings were soft; its AI roadmap is harder to read

Today's merged reading queue has one new analysis worth opening. Stratechery's public lede says Meta's earnings were "a bit disappointing," while the company's future AI-product promises were more disconcerting. The page is subscriber-only beyond that lede, so the summary below stays with the public framing and the chart visible on the page. 1

Advertising and AI

Stratechery: Meta's earnings and the timing problem

Published August 3, 2026. The public page gives three signals:

The quarter was weak, but not catastrophic. Stratechery's own wording is measured: Meta's earnings were "a bit disappointing." 1
The forward story is the sharper concern. The same public lede says promises about future AI products were "more disconcerting" than the earnings result itself. That makes timing the issue: the spending and promise arrive before the product payoff is visible. 1
The existing ad engine is still the financial baseline. The public chart compares ad-revenue growth, impressions, price-per-ad, and DAPs over time. Its latest points show ad-revenue growth above impressions, while price-per-ad and DAPs sit much lower. The gap is the useful signal for anyone trying to judge how much room Meta has to fund a still-unproven AI roadmap. 1

Line chart comparing Meta's ad-revenue growth, impressions, price-per-ad, and DAP growth over time — *The public chart is the article's clearest view of the "financial tail": four ad-business measures on one 2016–2027 timeline. 1*

Read the original Stratechery analysis

Lenny's archive

There is no new Lenny item to add today. The archive still tops out at the July 28 Product Pass expansion; the next post is dated July 21, and the July 28 item was already covered in the July 29 digest. 2 3

One thread to watch

The public evidence leaves a clean question for product and finance teams: can Meta turn its current advertising economics into visible AI products before the forward promises become a larger burden than the existing business can comfortably carry? Today's material does not answer that question—the Stratechery page withholds the full argument—but it does show why the question is about timing, not simply whether Meta is spending on AI. 1