Meta has virtually no AI users compared to OpenAI’s 900M, says Big Technology’s Kantrowitz

Alex Kantrowitz, founder of Big Technology and a CNBC contributor, delivered a pointed critique of Meta’s AI positioning this week, and the core of his argument comes down to one number: 900 million weekly active users on OpenAI’s ChatGPT.

“You have OpenAI building ChatGPT, which has 900 million weekly active users. Who knows, maybe it’s a billion right now. And how many intentional AI users does Meta have? Nothing close to that.”

That word “intentional” is doing a lot of work. Meta Platforms (NASDAQ:META | META Price Prediction) has its own headline number: Zuckerberg disclosed in Q1 2025 that Meta AI had “almost 1 billion monthly actives.” But Kantrowitz’s argument is that those two figures are not the same thing. Meta AI is embedded across Facebook, Instagram, WhatsApp, and Messenger. People encounter it whether they’re looking for it or not. ChatGPT users opened an app or navigated to a website specifically to use AI. One is a passive touchpoint; the other is a destination.

The distinction matters enormously for investors trying to assess whether Meta is actually competing in consumer AI or just borrowing the category’s language.

The Foundational Model Problem

Kantrowitz’s critique goes deeper than a user count. His argument is that a new category of consumer computing is being created right now, and Meta is not playing in it. The market has been waiting for Meta to deliver a foundational model competitive with OpenAI, and it hasn’t happened. Reports of the “Avocado” model rollout being pushed to May or later have only reinforced that concern. Prediction markets currently assign only a 10.5% probability to Meta’s “Mango” model releasing by end of March.

Kantrowitz’s conclusion: if Meta cannot make this work with its own internal technology, it is time to look outside for a solution. That’s a significant statement about a company spending $115-135 billion in capex in 2026 alone.

The financials are genuinely strong in isolation. Q4 2025 revenue came in at $59.89 billion, up nearly 24% year over year, with EPS of $8.88 beating estimates by over 8%. The advertising machine keeps printing. But operating margin compressed to 41% from 48% a year earlier as costs accelerated faster than revenue.

The stock reflects the uncertainty. META is down about 7% year to date, sitting well below its 52-week high of $795.06. Analyst consensus sits at $862.25, though that target assumes the AI investment eventually converts into durable competitive advantage.

Kantrowitz is essentially asking whether that conversion ever happens. If Meta is spending more than any company in history on AI infrastructure but still can’t produce a product that pulls users away from ChatGPT on purpose, the capex story starts to look less like a moat being built and more like a toll road with no cars on it. That’s the question Meta needs to answer in 2026, and right now the market doesn’t have a clear read on the answer.