Data & Targeting ScaleWide moat
Meta Platforms (META) — moat facet
Billions of daily actions teaching the models — more behavioral signal than almost anyone on earth.
Meta's advantage in advertising rests, at bottom, on an ocean of behavioral data gathered from billions of people using its apps for hours every single day. This is the raw material from which the whole machine is built, and its sheer volume is something almost no competitor can match. Every like, every pause on a video, every share and comment and follow is a signal, and the accumulation of trillions of such signals gives Meta a picture of human wants and behavior of extraordinary richness and resolution.
That data feeds the models that do the actual work of deciding which advertisement, and which piece of content, to place before which person. The sharper the targeting those models produce, the more an advertiser is willing to pay, and the more engaging the content they surface, the longer people stay — which generates still more data. Modern advertising is really a contest of prediction, and prediction is a game that rewards whoever has the most and the best material to learn from. Meta, by this measure, is very rich indeed.
The scale of engagement is itself a barrier no challenger can quickly surmount. The billions of hours people spend on Meta's apps each day are not merely a monetization opportunity; they are the very source of the data advantage, and a rival cannot manufacture that engagement on demand. To match Meta's data, a competitor would first have to match Meta's attention — and to match its attention, it would have to overcome the network effects that keep the attention where it is. The advantages compound and interlock.
There is a real and growing headwind here that an honest appraisal cannot ignore, and it is privacy. Regulators and platform owners have moved to restrict the tracking of individuals across the internet, and some of those changes have genuinely chipped at the edges of Meta's targeting ability, forcing the company to adapt its methods and absorb some real cost. This is the one place where the data moat can be narrowed not by a competitor but by a rule, and it deserves watching.
Yet the underlying advantage remains formidable, because so much of Meta's richest signal comes from within its own apps — from what people do on Facebook and Instagram themselves — rather than from tracking them elsewhere. The company has responded to the privacy restrictions by leaning harder on artificial intelligence to make the most of the first-party data it still holds in abundance, and by all appearances it has weathered the shift better than the early alarm suggested. The moat is narrower than it was, but it is very far from filled in: the data still flows from 3.58 billion people a day, the count at the end of 2025.1
Widening. Meta observes an enormous stream of engagement — what billions of people watch, like, share, and linger on — and at that scale AI can extract signal smaller platforms simply can't, turning attention into precise targeting and ranking. The more people engage, the more the models learn; the better the models, the more engaging and better-targeted the experience. This data-and-AI loop is Meta's deepest advantage, and it keeps compounding. The one real headwind is privacy regulation, handled in its own facet.
What the targeting earns per user. Growth falling toward DAP growth would mean the data advantage has stopped raising prices.
Source: Meta Form 10-Q, quarter ended 30 June 2026 ↗- ReportedThe moat is narrower than it was, but it is very far from filled in: the data still flows from 3.58 billion people a day, the count at the end of 2025.Meta Platforms, Forms 10-K (FY2021-FY2025) — revenue $117B -> $201B; 2022 net income halved to $23B; Reality Labs losses ~$19.2B (2025), >$80B cumulative; dual-class structure — FY2021-FY2025 · publ. 2022-2026 · source ↗