Meta, the Anti-Utility Company

Among big tech, Meta is the one that faces the sharpest innovator’s dilemma. The reason is simple: its business model makes it an anti-utility company.

Meta’s entire business is selling ads through an engagement machine. In Q1 2026, it generated $56 billion in revenue, of which $55 billion came from advertising — roughly 98%. It is hard to find another big tech more concentrated in a single revenue stream.

There is nothing intrinsically wrong with advertising. Google derives 70% of its $110 billion quarterly revenue from ads. But it’s built upon a very different product philosophy. From the beginning, Google set out to win the market by building the most useful search engine — getting users from question to answer as quickly as possible. (I love em-dash and nobody can take it away from me.)

Famously, when Larry and Sergey tried to sell Google to Excite (a portal website) in 1999, Excite passed. Portals made money by keeping users on their own pages, so Excite wanted a less effective search engine that doesn’t send users away so quickly.

Meta is, in a way, a more sophisticated Excite. Facebook and Instagram are not primarily task-completion tools. They are attention loops. Meta makes more money the longer users stay inside them. When someone goes to Google, they usually have a goal: find a fact, navigate somewhere, solve a problem. When someone opens Instagram, the goal is often much fuzzier: kill time, be entertained, avoid boredom.

That is Meta’s first dilemma. A utility helps people accomplish their goals as quickly and cleanly as possible. But that means less time spent, and thus fewer ads shown. Instagram probably has one of the richest real-time databases of local businesses in the world: whether a coffee shop is open, what a restaurant’s food actually looks like, which boutique just restocked an item. But turning Instagram into something like Google Maps or Yelp would point users toward answers, decisions, and exits. Nobody doomscrolls Google Maps for hours a day.

When looking beyond ads, this dilemma becomes more obvious. Zuck said in a recent interview that Apple has a “lack of innovation” problem. This may be true, but Apple is a fundamentally different business. It will do well if its M6 and A20 chips continue to be industry-leading. Likewise, if Nvidia keeps producing faster GPUs than the competition, if TSMC keeps shrinking the size of its transistors, and if SpaceX keeps making rockets cheaper and Starlink faster, they will do well. Hardware companies just need to make their hardware (objectively) better. Unlike Meta, their products are intrinsically useful.

The second dilemma is that advertising is too profitable. Simplified, Meta’s profit engine is:

profit = time spent × monetization per unit of time

Ads are extraordinarily attractive because they monetize attention without forcing Meta to solve messy real-world problems. Utility is harder. Helping people buy things, move things, resolve disputes, or complete transactions means higher costs and risks, more operational complexity, and lower margins. E-commerce is an obvious example: delivery, inventory, returns, fraud, customer support, and merchant reliability are all expensive. Meta would rather let merchants buy ads on Instagram and then push the hard parts to Shopify, DoorDash, or the merchant itself.

The engagement-machine model has served Meta exceptionally well for two decades. But in the long term, technology progresses by becoming more useful, not merely more addictive. That is why AI creates such a hard problem for Meta. The most valuable AI products are utilities: they answer, summarize, plan, search, transact, code, book, buy, and decide. They collapse time-to-outcome. They do not naturally maximize time-on-feed.

Meta can use AI to improve ad targeting, ranking, and recommendations, making its engagement machine even more effective. But the real question is whether it is willing to use AI to make its products genuinely more useful, which can hurt 98% of its business model.

That is the innovator’s dilemma of the anti-utility company: the next great technology wants to help users do things faster, while Meta’s core business still depends on keeping them around.