It's a strange setup for Meta Platforms, Inc.
That gap is where the story is.
The Case For Meta: AI Is Already Paying Off
Rosenblatt analyst Barton Crockett is blunt: Meta's AI spend is "justifying higher capex."
The data backs that up. Revenue grew 33% year-over-year to $56.3 billion, with AI helping drive engagement, ad impressions, and pricing. Importantly, that growth is arriving at roughly 60% EBITDA margins, pointing to what Crockett calls a "healthy ROIC" on those investments.
That's where the comparison gets uncomfortable.
Amazon.com, Inc's
In simple terms: Meta's AI is working. The market just isn't fully buying it.
The Case Against: Investors Still Want Proof
That skepticism shows up clearly in Goldman Sachs' take.
Analyst Eric Sheridan acknowledges that Meta's core business is "outgrowing the industry."
Ad revenue is accelerating, engagement is improving, and AI is clearly driving better outcomes across the platform-from content recommendations to ad conversion.
But that's not the issue.
The issue is visibility.
Sheridan points out that investors are still looking for clearer evidence of how Meta's AI investments translate into durable, long-term returns. In other words, the growth is there, but the line of sight into the payoff isn't.
And that matters more now than it did a year ago.
Execution Vs Perception
Put the two views together, and a single thesis emerges:
Meta is executing-but the narrative hasn't caught up.
AI is already lifting the business. It's improving engagement, monetization, and growth rates beyond pre-AI levels. But unlike cloud at Amazon or Microsoft, where revenue linkage is more direct, Meta's AI payoff is spread across multiple surfaces-and harder to model.
That makes investors cautious. Because in this phase of the AI cycle, it's no longer enough to spend-or even to grow. The returns need to be visible.
Until they are, Meta may continue to face a paradox: delivering results strong enough to justify its capex-while still having to prove that they matter.
