Zuckerberg's metaverse couldn't go far without legs

BarcelonaMark Zuckerberg has put an end to the virtual reality version of Horizon Worlds, the platform that was supposed to be the gateway to his metaverse. He announced it last Tuesday on Meta's official forums and in emails to users. It's no surprise: it's the formal liquidation of a project that had been agonizing for years and never surpassed 300,000 monthly active users; a ridiculous figure for a company that hosts 3 billion humans on its platforms.

The closure will be progressive: on March 31st it will disappear from the Quest Store and on June 15th it will be definitively erased from the headsets, eliminating all community-created worlds. The mobile version will continue to function, but surviving in a market dominated by Roblox and Fortnite doesn't seem very promising.

The name change, legs, and other distractions

To understand the failure of Horizon Worlds, we need to go back to October 2021, when Zuckerberg pompously presented the rebranded Meta Platforms and proclaimed that the metaverse would reach "a billion people." At that time, Facebook and Instagram already carried a regrettable reputation: scandals of political manipulation, misuse of personal data, inaction in the face of child harassment. The strategy was clear: change the name of the entire company and divert attention towards a grandiose vision of the digital future.

Here we already pointed out, that November 2021, that Facebook's forward flight had three engines: the reputation crisis derived from its toxic practices; the aging of its user base –accelerated by the success of TikTok–; and the desire to escape the dependence on Apple and Google, which dictated the rules and took a commission from each transaction. In its own metaverse, Zuckerberg could apply its rules without tolls. Under the cover of a supposed bright technological future, the goal was business.

Many people took Zuckerberg literally. Dozens of companies rushed to appoint "metaverse directors," and McKinsey assured in 2022 that the metaverse would generate up to five trillion dollars in value by 2030. That report has aged terribly: the vision never translated into a product worthy of such ambition. The image that best summarizes it was inadvertently leaked by Meta itself: the virtual Zuckerberg walking around with the lower part of his body non-existent. The avatars had nothing from the waist down. Literally, they had no legs. A few months later, Meta announced that it would equip them with lower limbs – and admitted that the demonstration video had been recorded with conventional motion capture, not within the metaverse.

In our 2022 technology summary, we wrote that "the metaverse has very short legs – in fact, it has none." It's almost a shame we got it so right.

Horizon Worlds figures document the extent of the failure. In February 2022, Meta celebrated the platform's peak: 300,000 monthly active users. Eight months later, internal documents leaked to the Wall Street Journal showed they had fallen below 200,000, that most did not return after the first month, and that 91% of created worlds had never received more than 50 visits. From 2023, Meta stopped publishing figures. A YouTuber who spent a week on the platform found only 900 active users at any given time. Let's compare this to Roblox's 144 million daily active users.

83.5 billion dollars thrown away

The Reality Labs division has been accumulating ever-increasing losses since 2020: over 83.5 billion dollars in six years, against an accumulated turnover of less than 12 billion –generated mainly by the sale of Quest headsets–. Adding the purchase of Oculus in 2014 and other previous investments, the total bill exceeds 100 billion.

Meta has been able to absorb these losses thanks to the profits from Facebook, Instagram and WhatsApp –83.3 billion in 2025–. But when Bloomberg leaked last December that the company was preparing to cut the metaverse budget by 30%, shares rose 4%: the market was once again telling Zuckerberg what he didn't want to hear.

Just over two months ago, Meta laid off 1,500 Reality Labs employees –10% of the division–, closed three video game studios, canceled a Batman sequel, and suspended the fitness app Supernatural, which it had bought for 400 million. CTO Andrew Bosworth summarized it in Davos: "VR is growing less rapidly than we expected."

The coup de grâce of generative AI

In November 2022, OpenAI launched ChatGPT and the industry changed suddenly. The metaverse, which already had serious adoption problems, was relegated in favor of a technology that offered immediate and profitable results. Meta grasped this immediately: in February 2023 they published Llama 1, and by the end of the year their shares had risen by 194%. In the third quarter of 2024, executives did not utter the word "metaverse" even once in the entire earnings conference.

The change in direction is total. This year, Meta plans to allocate more than 120 billion dollars to AI infrastructure, practically double that of last year, and Vishal Shah – the former head of the metaverse – now leads the new Superintelligence Labs unit. Generative AI did not directly kill the metaverse because it was already dying on its own, but it has provided Zuckerberg with a much more convincing future narrative and a product that people actually use.

The modest success of connected glasses

Amidst so much failure, Meta can point to a hardware success that no one would have predicted: the Ray-Ban Meta smart glasses. The first generation, in 2021, was a flop, but the second, presented in 2023, tripled sales during 2025 to exceed 7 million units. The explanation is simple: they don't require any change in behavior; just wear them and talk to them to take photos, listen to music, or chat with Meta AI. However, the glasses have cameras, microphones, and speakers, and Meta is not exactly known for its restraint when it comes to exploiting user data. On these pages, we have been reviewing the contradictions of the initiative: the rebranding of Facebook as a public relations operation; legless avatars as a metaphor for an unfinished product; consumer metaverses as clumsy solutions that looked for problems to solve. Deep down, the metaverse's sole aim was to create a closed environment where Meta could set its own rules without paying commissions to Apple and Google, disguised as a vision of the future. Generative AI is not necessarily more altruistic, but it is infinitely more useful, and people are using it.

Zuckerberg's metaverse couldn't go far without legs. And in the end, it hasn't gone very far with them either.