Meta keeps throwing cash at Reality Labs with no end in sight

Meta’s Reality Labs division has been a money pit for years, and the latest earnings report doesn’t change that picture. The segment lost another $4.5 billion last quarter, bringing its total operating loss since 2020 to well over $50 billion. That’s not a typo.

I’ve been watching this space since the Oculus Rift days, and even I find those numbers staggering. Mark Zuckerberg has made it clear he sees AR/VR as the next computing platform, but at what point does “long-term bet” become “fiscal black hole”?

The thing is, Reality Labs isn’t just burning cash on hardware. They’re funding entire ecosystems—Horizon Worlds, Quest software, developer grants, and now AI research that’s increasingly tied to those efforts. The AI angle is interesting because it’s a double-edged sword: Meta needs massive compute infrastructure for its generative AI ambitions, and that doesn’t come cheap. Their capital expenditures jumped to $9.2 billion this quarter, mostly on data centers and GPUs.

So here’s the math problem: Reality Labs loses billions, AI spending adds billions more, and the core advertising business—which still prints money—has to carry all of it. Ads revenue grew 18% year-over-year, which is healthy, but not enough to make these side bets look anything but reckless.

I’m not saying Meta should kill Reality Labs. The Quest 3 is genuinely good hardware, and the Passthrough AR features are impressive for what they are. But the company has been promising a breakthrough AR glasses product for years, and every time they push the timeline back. Meanwhile, Apple’s Vision Pro is eating the high-end mindshare, and even if it’s not a sales hit yet, it’s setting the standard.

What bothers me is the lack of discipline. Meta could be funneling some of that cash into more practical AI applications—better recommendation algorithms, smarter ad targeting, tools for creators—but instead they’re doubling down on a vision that might not materialize for another decade. Zuckerberg has the luxury of controlling the board, but shareholders are getting restless.

The real test will come if ad revenue growth slows. Right now, Meta’s core business is strong enough to absorb the losses, but that won’t last forever. If the economy dips or competition from TikTok and others heats up, those Reality Labs losses will start to look a lot less tolerable.

I’d love to be wrong about this. I want AR glasses that don’t look ridiculous and actually work. But throwing money at a problem doesn’t solve it—it just delays the reckoning. Meta needs to show some real progress, not just bigger losses and prettier demos.