Recently, there’s a major tech event worth pondering. Meta just announced that it will shut down the VR version of Horizon Worlds on June 15th this year. This once hyped-up metaverse dream is now officially declared dead. Honestly, looking at this decision, I can only describe Meta’s attitude towards its past choices with a “middle finger meme.”

Recalling the original grand vision Zuckerberg painted—to hold meetings, entertain, and achieve digital freedom in virtual space—Reality Labs has burned nearly $80 billion over the years, only to end up with an abandoned platform that no one uses. Can you imagine? Spending so much money just to get extremely low daily active user numbers. From a financial report perspective, it’s basically a bottomless money pit.

Analyzing the failure of Horizon Worlds, there are actually three fatal flaws. First, the application scenario simply doesn’t hold up. Video conferencing software already solves 90% of remote work needs, so why would people wear bulky VR headsets to enter a virtual space that only shows their upper body? That’s fundamentally a false demand. Second, hardware costs are too high. The expansion of Quest devices has fallen far short of expectations, preventing network effects from forming. Lastly, the content ecosystem is a mess. Platforms like VRChat and Roblox already have strong decentralized creator communities, but Meta tried to build a community through top-down resource infusion, which of course failed.

But the real highlight of this event isn’t Horizon Worlds’ failure itself, but Meta’s strategic shift as a whole. Zuckerberg has long been fixated on the development of general artificial intelligence. After the open-source Llama model achieved overwhelming success within the developer community, Meta realized that dominating AI infrastructure has a much higher chance of success than struggling to sustain a virtual paradise.

So you see, from passively guiding users into virtual space to actively controlling the next-generation computing power, this is Meta’s key pivot. The resources once invested in Horizon Worlds are now pouring into large language model iterations, data center expansion, and custom AI chip development.

Hardware strategy has also shifted. Over the past five years, Meta poured heavy investments into fully enclosed VR headsets, but market response has been clear: modern people’s desire to “escape reality” is far less than their reliance on “augmented reality.” So now, Meta’s hardware focus has shifted to lightweight smart glasses like Ray-Ban Meta. These devices don’t force users out of their physical environment but seamlessly integrate with the real world through voice assistants and image recognition. This “AI-driven, hardware-assisted” approach not only reduces development costs but also continuously collects firsthand visual and auditory data streams to improve the underlying AI models.

In short, Meta is willing to endure the $80 billion sunk cost and withdraw because they’ve already seen the next battlefield they must win—and can win. The era of the metaverse is finally coming to an end.