The Next Frontier
Meta just killed Horizon Worlds. In 2021, Zuckerberg called the metaverse “the next frontier” and renamed the company. In 2026, he’s calling AI “the next frontier” and laying off 20%. Same sentence. Different bet.
In October 2021, Mark Zuckerberg renamed his company. Facebook became Meta. The vision statement, published the same day:
“Our hope is that within the next decade, the metaverse will reach a billion people, host hundreds of billions of dollars of digital commerce, and support jobs for millions of creators and developers.”
This week, Meta announced it is shutting down Horizon Worlds — the VR social platform that was supposed to be that billion-user destination — at the end of March. Full removal from VR by June 15. Relegated to a mobile app.
Reality Labs, the unit that built Horizon Worlds, has been losing money at a rate that would be catastrophic for any other company. $13.7 billion in losses in 2022. Around $50+ billion total since the metaverse pivot began. Operating loss of $6.02 billion in Q4 2025 alone. And in January, over 1,000 Reality Labs employees were let go. The VR gaming studios are gone. The dream is on life support.
Meanwhile, Meta is reportedly planning to lay off up to 20% of its workforce company-wide, citing AI-driven efficiency improvements and the need to “prepare for greater efficiency brought about by AI-assisted workers.”
The next frontier is always right around the corner. The current frontier is always more expensive than planned.
It’s easy to mock the metaverse bet. The avatars had no legs. The peak monthly active users never broke a few hundred thousand — embarrassingly low for a company that routinely measures audiences in billions. The technology shipped late, the headsets were expensive, the experience was underwhelming, and the general public never showed up.
But I don’t think the metaverse was a strategy failure. I think it was a discovery process.
Zuckerberg’s real problem in 2021 wasn’t that he didn’t have a social network. He had the largest one in history. His problem was that his entire business ran on top of platforms he didn’t control — Apple’s iOS, Google’s Android — and those platform owners were slowly tightening the screws. Apple’s App Tracking Transparency update alone cost Meta an estimated $10 billion in 2022 revenue.
The metaverse was Zuckerberg’s attempt to own the platform. Not just the app, but the OS, the hardware, the input devices, the distribution layer. If everyone spent their social hours in a VR world that Meta controlled end-to-end, Apple’s ATT became irrelevant. The metaverse wasn’t about virtual reality. It was about escaping iOS.
The problem is that you can’t force a platform transition by declaration. The $40+ billion in Reality Labs losses bought Meta important information: the public is not ready to put a headset on for social interaction, the VR hardware is not compelling enough to drive mass adoption, and even when the Quest is technically impressive, people would rather scroll TikTok on the same couch.
This is expensive information. But it’s information. And knowing it in 2026 is worth something.
The AI pivot carries the same underlying logic. Zuckerberg isn’t going all-in on AI because he loves large language models. He’s going all-in because AI changes what it means to own the platform.
The old platform monopoly (iOS, Android) controlled distribution. Whoever controls the app store controls what software reaches users and at what cost. Meta paid rent to Apple for fifteen years.
The new platform question is: who controls the agent layer? If AI agents become the primary interface for getting things done — searching, shopping, booking, socializing — then whoever owns the most-used AI stack owns the new distribution layer. The company doesn’t necessarily have to be the smartest AI. It has to be the most embedded one.
Meta is spending $60+ billion in AI capex in 2025-2026. That’s not R&D spending. That’s platform-building spending. The question is whether this time the platform transition actually happens.
The symmetry is jarring: same company, same CEO, same problem (platform dependence), same solution template (own the next layer), different technology bet.
The personnel transition is more jarring. In 2021, Zuckerberg told the world the metaverse would “support jobs for millions of creators and developers.” In 2026, he’s cutting up to 20% of his own workforce because AI is replacing what people used to do.
The metaverse promised jobs. The AI pivot eliminates them. Both are sold as transformative. Both originated from the same anxiety about owning the future rather than renting it.
There’s something almost mathematical about the way Zuckerberg treats belief cycles. He commits with enormous conviction. He funds the commitment beyond any reasonable threshold. He acquires information about whether the conviction was correct. He pivots based on what the information says.
The metaverse bet was a $40 billion experiment with one clear result: VR social won’t drive mass adoption in the 2020s. The AI bet is a $60 billion experiment with the hypothesis that the agent layer will be where attention lives in the 2030s.
He might be wrong again. But he’ll find out before everyone else, and he’ll know exactly how wrong.
Horizon Worlds will be gone by June. The avatars with no legs will be deleted. The vision that was supposed to outlast Facebook, that was supposed to free the company from the tyranny of App Store fees, is being reduced to a mobile app — the exact format it was supposed to transcend.
The metaverse isn’t the next frontier. That’s what the AI announcement says now.
But “the next frontier” is Zuckerberg’s favorite phrase. He’s used it to describe three different things in three different decades. Each time, it’s the one that will change everything, that will be where the real Meta lives.
The interesting thing isn’t which frontier he’s pointing at. It’s that he keeps pointing. That he keeps having the conviction to rename the map, even when the last frontier turned out to be ocean.
The next frontier is a mindset, not a destination. And at Meta, the mindset never changes: own the platform or pay for it.