No experiment of this scale ends without leaving something behind — even if that something is primarily a lesson. Meta is shutting down Horizon Worlds on VR, pulling it from the Quest store in March and ending all VR access on June 15. Mark Zuckerberg’s metaverse cost close to $80 billion and reached far fewer people than promised, but the experience will shape Meta’s AI strategy in ways that are only beginning to emerge.
The 2021 rebrand from Facebook to Meta was a declaration of strategic intent. Zuckerberg had concluded that the future of human communication would be immersive, spatial, and avatar-driven. He invested the company’s brand and billions of dollars in making that future real. The ambition was clear; the path to realizing it proved far more complicated than anticipated.
Horizon Worlds was built with care and updated with persistence. But it could not escape a fundamental challenge: VR headsets remained niche consumer devices, limiting the platform’s potential user base from the start. Without mass device adoption, Horizon Worlds could never achieve the scale of social interaction that made platforms like Facebook valuable. Its few hundred thousand monthly users represented a ceiling, not a floor.
Reality Labs posted close to $80 billion in losses over four years of development, representing one of the most sustained single-product investments in tech history. In early 2025, more than 1,000 Reality Labs employees were let go, and Meta formally began its shift toward AI and wearables — two categories where it sees more immediate commercial opportunity and competitive urgency.
The lessons of the metaverse — about market timing, consumer readiness, and the dangers of anchoring a company’s identity to a single technology bet — will inform how Meta approaches AI. Zuckerberg will move more carefully, test assumptions more rigorously, and perhaps resist the temptation to rename the company Anthropic. The metaverse was expensive, but its lessons may prove valuable.
