⚠ The Attention RecessionModerate threat

Engagement is a zero-sum contest for a fixed resource — there are only so many waking hours — and Meta must defend its share against an intensifying field. TikTok captured enormous amounts of time from young users; YouTube dominates long and short video; gaming, streaming, and a hundred other apps all compete for the same finite attention. Every hour won by a rival is an hour of data and advertising inventory Meta does not get.

The danger is that Meta can keep the social graph while losing the engagement, and it is engagement, not membership, that the machine runs on. If the most valuable users — the young, the highly-engaged — give more of their time to competing formats, Meta's apps risk becoming accounts people keep but no longer live in, which starves the data-and-advertising flywheel even if no one formally leaves.

Against it stands Meta's demonstrated ability to win attention back by matching the winning format — Reels answered TikTok and now drives¹ a large and growing share of time on Instagram and Facebook — and by using AI to make its feeds more engaging. Daily users were still rising in 2026, up 3% to 3.60 billion in June, even amid the competition.²

This one is moderate and permanent. The battle for attention never ends, and a format Meta cannot copy could genuinely erode the engagement its whole machine depends on — but Meta has repeatedly defended its share by matching rivals and out-ranking them, and its engagement has so far held and even grown.