$META calls — AI ad personalization is the margin story nobody is modeling correctly

Everyone knows Meta has a big ad business. What gets underpriced is how much their AI personalization is actually changing the unit economics of that business.

When their models get better at matching ads to intent, advertisers get more ROI per dollar. That means they can charge more without volume falling off. It is not just "more ads" — it is the same inventory becoming structurally more valuable. Reels monetization proved this out. The gap between Reels engagement rates and ad load was huge at launch and has been closing faster than expected.

The Reality Labs losses are real but finite. The glasses and wearables category is still early, but the underlying hardware business is not a bottomless pit — it has a defined roadmap and the burn is capped. Most bears are modeling RL losses as permanent drag forever, which is not how product cycles work.

What I keep coming back to is the network effect compounding. Instagram, WhatsApp, and Facebook are not going to lose billions of users simultaneously. The engagement floors are durable, and AI-driven feed improvements actually increase DAUs over time. More users, better ad targeting, pricing power — that is a formula that works.

Sitting on calls here. The market has been treating any AI capex headline as a negative, but the ROI evidence from the ad business says otherwise. Tariff noise hit the stock in March and it has not fully recovered. That is the entry.