ON TUESDAY, Meta Platforms announced that users of its Quest VR headsets would lose access to Horizon Worlds — the cartoon-avatar social platform that was once the centerpiece of Mark Zuckerberg's grandest corporate pivot — effective June 15. A mobile app version will survive, though calling that the "metaverse" requires a generosity of definition that would make even a venture capitalist blush.
The announcement, while small in practical terms, is symbolically enormous. On October 28, 2021, Zuckerberg renamed his entire company — the most used social network in human history — after the concept. "Over time, I hope that we are seen as a metaverse company," he told the world. Facebook would become Meta. The blue thumbs-up would yield to an infinity symbol. The future would be three-dimensional, immersive, and populated by legless avatars attending virtual concerts. The bill for this conviction has since climbed to roughly $90 billion in cumulative operating losses at Reality Labs over seven years. Annual losses escalated reliably — $6.6 billion in 2020, $10.2 billion in 2021, $13.7 billion in 2022, $16.1 billion in 2023, $17.7 billion in 2024 — and $19.2 billion in 2025. By any accounting standard, it was one of the most expensive cul-de-sacs in corporate history.
Legs to stand on (finally, none)
But what makes this epitaph particularly wry is not the money — Meta prints cash from its advertising business with the ease of a government mint — but the product itself. Horizon Worlds peaked at roughly 300,000 monthly users in early 2022. By October of that year, that figure had slid below 200,000, and the company quietly halved its year-end target from 500,000 to 280,000. Only 9% of user-created worlds were ever visited by more than 50 people. One YouTuber who spent a week inside the platform in 2023 reported encountering roughly 900 daily users — fewer than a mid-tier Roblox server, in a product that had consumed more capital than the GDP of most Caribbean nations.
The failure was not, strictly speaking, technological. VR headsets improved markedly during this period; the Quest 3 is by most accounts a capable device. The failure was anthropological. People did not want to attend meetings as cartoon torsos. They did not yearn to visit virtual comedy clubs with strangers. The metaverse solved a problem that did not exist for a consumer base that was perfectly content scrolling Instagram in two dimensions while lying on a couch — a behavior Meta's own advertising division was rather expertly monetizing.
Zuckerberg, to his credit, reckons with sunk costs more ruthlessly than most founder-CEOs. Reality Labs began shedding headcount in January, with roughly 1,000 jobs eliminated from the division. CTO Andrew Bosworth, who leads the unit, told staff the company would pivot toward mobile experiences — an implicit confession that the VR-first thesis had stalled. Resources have since flowed toward the Ray-Ban Meta smart glasses and, more broadly, toward the artificial intelligence arms race where Meta's Llama models have made the company a formidable open-source contender.
The pivot is instructive. The metaverse was, at bottom, a platform play: build the next computing paradigm, own the hardware and software stack, escape dependence on Apple and Google's app stores. AI offers a version of the same strategic logic — a new platform, a new distribution layer — but with the crucial advantage of being something people actually want to use. Meta's AI assistant is now embedded across its apps; its open-weight models are downloaded millions of times. The contrast with Horizon Worlds, where the challenge was persuading anyone to show up at all, could hardly be sharper.
None of this means the $90 billion was entirely wasted. The hardware R&D has produced the Ray-Ban glasses, which have found a genuine, if niche, consumer audience. Spatial computing expertise may yet prove valuable as augmented reality matures. And Meta can afford the tuition: the company generated north of $160 billion in advertising revenue in 2024, making Reality Labs' losses a painful but not existential line item.
Still, the lesson is worth cataloguing. The metaverse's failure was not that the technology was bad or the vision impossible; it was that Zuckerberg mistook a supply-side insight for demand. He saw where computing could go and assumed consumers would follow. They did not. They rarely do on command, no matter how much money is spent building the road. The infinity symbol, it turns out, was less a logo than a warning about how long it takes to reach a destination no one asked to visit. ■
For more, join 75,000 subscribers getting tech's favorite brief here
