What the Metaverse Got Wrong and What Comes Next
Meta poured more than $80 billion into a virtual world nobody wanted to visit. Here is why the bet failed and how spatial computing is quietly succeeding where the metaverse could not.
The metaverse, as Silicon Valley pitched it in 2021, promised a persistent 3D internet where billions of people would work, socialize, and shop through virtual avatars.
Trending Now!!:
Five years later, the flagship product built to prove that vision, Meta’s Horizon Worlds, has been stripped down to a mobile app after burning through tens of billions of dollars, while the technology it was supposed to replace, ordinary screens and video calls, still does the job better for most people.
What comes next is not a bigger metaverse. It is a narrower, more useful one: enterprise-first spatial computing built around digital twins, training simulations, and AI-assisted headsets rather than social avatars in a shared virtual mall.
The Bet That Defined a Company
In October 2021, Mark Zuckerberg renamed Facebook to Meta. He described the metaverse as “the next frontier,” a virtual layer of the internet he predicted would reach a billion people and generate hundreds of billions of dollars in commerce within a decade.
Zuckerberg described the metaverse as “the next frontier,” predicting it would reach a billion people and generate hundreds of billions in digital commerce within a decade Horizon Worlds, the VR social platform built to anchor that vision, launched in December 2021 as the primary destination users would reach through Meta’s line of Quest headsets.
The numbers never came close. Horizon Worlds peaked at approximately 300,000 monthly active users in February 2022, then declined to fewer than 200,000 by October 2022, a shortfall from an internal target that had already been revised down twice, first from 500,000 users to 280,000.
By October 2022, just months after its public launch, Horizon Worlds had fewer than 200,000 monthly active users, well short of an internal target of 280,000, which itself had been revised down from an original goal of 500,000. By 2024, an independent investigation put daily active users as low as 900, with most people never returning after their first month.
Only 9 percent of user-created worlds ever attracted more than 50 visitors. By 2024, an independent investigation found as few as 900 daily active users. Most users never returned after their first month, and only 9% of user-created worlds ever attracted more than 50 visitors.
The financial exposure behind those numbers is what makes Horizon Worlds a genuine industry case study rather than a footnote. Reality Labs, Meta’s division for VR and metaverse development, has posted an operating loss every year since 2020, with the cumulative total reaching $83.6 billion through 2025.
Reality Labs, the Meta division responsible for VR and metaverse development, has posted operating losses every year since 2020, with deficits growing each year. The cumulative total reached $83.6 billion through 2025. In the fourth quarter of 2024 alone, the division lost $6.02 billion on revenue equal to roughly 1 percent of Meta’s total. In the fourth quarter of 2024 alone, the division posted an operating loss of $6.02 billion on revenue that represented roughly 1% of Meta’s total.
Where the Product Failed Before the Strategy Did
Industry analysts tend to frame the metaverse’s collapse as a strategic miscalculation. That framing understates how badly the product itself misjudged its own users.
Horizon Worlds shipped with avatars that had no legs. This detail became an easy punchline but reflected something deeper: the platform launched before its core interaction model was finished, and the negative first impression it created never fully reversed.
The deeper problem was structural. Horizon Worlds required a $300-plus VR headset to access a social network, but the value proposition of a social network depends on people already being there. Regional availability made this worse in the early years, limiting Horizon Worlds to a handful of countries at launch and shrinking the addressable audience before the hardware barrier was even considered.
Expansion of the platform was limited due to small regional availability, with Horizon Worlds available in only seven countries, and the necessity of a compatible VR headset. A social platform that gates itself behind expensive, uncomfortable hardware is not competing with other social platforms. It is competing with the phone already in someone’s pocket, and it loses that fight by default.
The most useful counterexample sits directly next to Horizon Worlds’ failure. VR gaming has not struggled the way VR social platforms have. Titles built around a clear, repeatable activity, such as Beat Saber and Half-Life Alyx, retained committed audiences because they solved a problem headsets are actually good at solving: immersive, single-purpose entertainment. Video calls, screen sharing, and text chat had already solved remote human connection well enough that a clunky, isolating VR alternative offered no real upgrade for most people, only a heavier one.
The Reversal Nobody Predicted Would Come So Fast
By March 2026, Meta confirmed what markets had anticipated for years: it was winding down the VR version of Horizon Worlds. Individual worlds began disappearing from the Meta Horizon Store, Worlds and Events were set to leave the Quest Store by March 31, and the app itself was scheduled for removal from Quest headsets entirely on June 15, shifting the product to mobile only.
Under the plan, Horizon Worlds and Events would disappear from the Quest Store by 31 March 2026, several flagship worlds would no longer be available in VR, and the Horizon Worlds app itself would be removed from Quest on 15 June 2026, ending VR access to Worlds altogether.
What happened next is arguably more revealing than the shutdown itself. Facing user backlash, Meta reversed course within days. Chief Technology Officer Andrew Bosworth confirmed in an Instagram Stories Q&A that Horizon Worlds would remain available in VR after all.
In an Instagram Stories Q&A, CTO Andrew Bosworth said Horizon Worlds would remain available in VR after user backlash. The reversal did not change the underlying trajectory. Meta had already confirmed that mobile, not VR, was the platform’s future, a decision that landed as Reality Labs’ cumulative losses hit $83.6 billion.
Meta reversed plans to shut down VR Horizon Worlds after user backlash, but confirmed mobile as its future focus as Reality Labs losses hit $83.6 billion.
The capital reallocation tells the real story. Meta is directing $115 billion to $135 billion toward AI infrastructure in 2026, up from $72 billion in 2025, and has formed a new Superintelligence Labs division under Zuckerberg’s direct oversight, while Reality Labs absorbs budget cuts of up to 30 percent alongside studio closures.
Meta is now investing $115-135 billion in AI infrastructure for 2026, up from $72 billion in 2025, forming a new Superintelligence Labs division under Zuckerberg’s direct oversight, while Reality Labs faces 30% budget cuts, studio closures, and layoffs. A company does not shift capital at that scale unless the previous bet has definitively failed to clear its return threshold.
What Was Actually Working the Whole Time
The metaverse narrative obscured an inconvenient fact for Meta: virtual worlds were succeeding elsewhere, just not under that name and not built around corporate ambition. Roblox commands roughly 214 million monthly users and Fortnite attracts around 236 million, both proving that persistent virtual worlds succeed when they grow around genuine player motivation rather than a top-down platform vision.
Roblox commands 214 million monthly users while Fortnite attracts 236 million, proving that virtual worlds succeed when built around genuine user motivations rather than corporate visions.
Neither asked users to abandon a keyboard and a screen for an avatar in goggles. Both simply extended existing gaming behaviour into more persistent, social, creator-driven spaces, which is a fundamentally different product bet than Horizon Worlds ever made.
Enterprise adoption followed a similar pattern of quiet success beneath the consumer failure headlines. Roughly 23 percent of Fortune 500 companies now allocate dedicated spatial computing budgets averaging $2.7 million annually.
Half of Fortune 100 companies have purchased Apple Vision Pro units for training and design applications.23% of Fortune 500 companies allocate dedicated metaverse budgets averaging $2.7 million annually, with half of Fortune 100 purchasing Apple Vision Pro units for spatial computing applications. The clearest evidence of measurable return comes from manufacturing.
BMW has used Siemens-NVIDIA digital twin partnerships to cut manufacturing design errors by 40 percent and reduce time-to-market by 28 percent, aided by simulation speeds up to 30 times faster than prior methods.BMW reduces manufacturing design errors by 40% and cuts time-to-market 28% through Siemens-NVIDIA partnerships enabling 30x simulation speed improvements and real-time factory optimization
Healthcare has produced some of the most concrete outcomes in the category. VA pilot centers report VR therapy reducing PTSD recovery time by 30 percent, and Johns Hopkins has conducted remote surgeries with more than 90 percent accuracy using immersive tools.VR therapy reduces PTSD recovery time by 30% across VA pilot centers.
At the same time, Johns Hopkins conducts remote surgeries with 90%+ accuracy, demonstrating clinical applications with measurable patient outcomes. These are the use cases that never needed the word metaverse. They needed a specific tool for a specific, high-stakes task, which is precisely the category Horizon Worlds never occupied.
Spatial Computing: The Rebrand That Reflects a Real Shift
Apple never used the word metaverse. When it launched Vision Pro, the company positioned spatial computing as the umbrella term for AR and VR, deliberately distancing the product from the social-world baggage Meta had accumulated, and the industry has largely followed Apple’s lead in adopting that language.
With the launch of the Apple Vision Pro, Apple officially embraced Spatial Computing as the preferred category term for their Mixed Reality headset, helping to standardize a common terminology of the industry. The distinction is not just branding. It reflects a genuinely different design philosophy: instead of replacing a user’s environment with a virtual one, spatial computing overlays digital content onto physical space, which is a far better fit for enterprise workflows than for casual social interaction.
Apple Vision Pro 2, built on the M5 chip with twice the on-device AI inference speed of its predecessor, launched at $2,499 in February 2026, a meaningful price cut from the original $3,499 model, alongside visionOS 26’s spatial scenes framework for persistent, context-aware 3D environments.
Apple Vision Pro 2, powered by the M5 chip with 2x faster on-device AI inference, dropped to $2,499 at launch in February 2026, and visionOS 26 introduced spatial scenes, a framework that lets developers build persistent, context-aware 3D environments that blend with physical workspaces.
The industrial use case has become the category’s clearest proof point: a digital twin displayed on a flat screen is just a dashboard, but the same twin rendered in spatial computing becomes a walkable control room, with real-time sensor data floating above the physical equipment it monitors and remote collaborators appearing as spatial presences inside the same environment.
A digital twin on a 2D screen is a dashboard. A digital twin in spatial computing is a control room you can walk through, letting operators walk through a virtual replica of the facility, see real-time sensor data floating above the equipment it monitors, and collaborate with remote team members who appear as spatial personas in the same twin
The market figures reflect that enterprise-first tilt. Spending on spatial computing services is growing at a 44 percent compound annual rate, faster than any other segment tracked in the category, with manufacturing, healthcare, and engineering leading adoption based on documented return on investment rather than novelty.
Enterprise spending on spatial computing services is growing at a 44% CAGR, faster than any other segment, and enterprise manufacturing, healthcare, and engineering are leading adoption based on ROI evidence.
Meta still holds the largest share of the XR headset market at roughly 50.8 percent, but the growth story has moved away from Meta’s original consumer vision toward challengers building specifically for business buyers. Meta holds approximately 50.8% of XR headset market share, with Apple, Samsung, and Google as the major challengers.
The Convergence Point: AI, Glasses, and a Chicken-and-Egg Problem That Persists
Vision Pro still faces the same structural obstacle that undercut Horizon Worlds, just in a different form. Developers hesitate to build fully spatial experiences without a large enough installed base, and consumers hesitate to buy a premium headset without enough spatial-native content to justify the price, a standoff analysts describe plainly as a chicken-and-egg problem.
Vision Pro faces the classic chicken-and-egg problem: developers will not build fully spatial experiences until there is a bigger consumer base to build for Enterprise buyers have sidestepped that trap by purchasing outcomes rather than platforms, paying for specific training, design, or simulation results rather than betting on a consumer ecosystem that may never fully materialize.
The technology that appears best positioned to break the standoff is not a headset at all. Meta has already shifted toward lightweight AR glasses distributed to developers, and infrastructure providers like NVIDIA have released cloud streaming tools built specifically to offload the heaviest spatial computing workloads away from the device itself.
Meta shipped its lightweight AR glasses to developers, and NVIDIA launched CloudXR 4.0 for streaming heavy spatial workloads from the cloud. That architecture matters because it removes the two biggest barriers that killed consumer VR adoption: device cost and processing power. Lighter glasses paired with cloud-rendered content, layered with AI systems capable of contextual reasoning about a user’s physical surroundings, is a far more plausible mass-market path than a standalone VR headset ever was.
What Comes Next
The next phase of this technology will not be defined by a single platform chasing a billion users. It will be defined by narrower, ROI-driven deployments that expand outward from proven use cases rather than a top-down social vision imposed on a market that never asked for it. Four shifts are already visible in how the category is rebuilding itself.
Enterprise adoption will continue to outpace consumer adoption for the foreseeable future, because manufacturing, healthcare, and engineering buyers can measure return in reduced errors, faster training, and fewer costly mistakes.
In contrast, consumer social use cases still cannot articulate a problem that a phone or a video call does not already solve adequately. Hardware will keep splitting into two distinct categories rather than converging into one universal device: heavier, high-fidelity headsets for professional and industrial work, and lighter AI-integrated glasses aimed at everyday consumer use.
AI will do more of the interpretive work that VR alone could never manage, turning static digital twins into systems that reason about the physical environment they represent rather than simply rendering it. And the word metaverse itself will keep fading from serious industry conversation, not because the underlying technology failed, but because the term became inseparable from a specific, overreaching vision that the market rejected on its merits.
The lesson for anyone building in this space is not that immersive technology failed. Enterprise applications, VR gaming, and AI-integrated glasses are all growing on legitimate demand and measurable results. What failed was the assumption that a company could manufacture demand for an entire alternate social reality simply by investing enough capital in it.
The next wave of spatial computing is succeeding precisely because it abandoned that assumption and started solving specific, verifiable problems instead.
What People Ask
