Spatial computing has quietly replaced the metaverse as the term that matters, and the numbers explain why. Meta was considering cuts of up to 30 percent to its Reality Labs budget for 2026, according to reporting confirmed by a Meta spokesperson, who said the company was “shifting some of our investment from metaverse towards AI glasses and wearables given the momentum there.” The unit that Mark Zuckerberg bet his company’s future on when he rebranded from Facebook to Meta in 2021, committing tens of billions of dollars to what he called the next frontier of computing, is in retreat.
Daily active users in crypto-metaverse platforms like Decentraland and The Sandbox dwindled into the hundreds at their nadir. Consumer metaverse platforms became ghost towns.
And yet, at exactly this moment of apparent collapse, the underlying technology that the metaverse was always meant to deploy is maturing, rebranding, finding enterprise applications, and positioning itself for a growth trajectory that has nothing to do with virtual real estate speculation. Understanding what actually happened to the metaverse, and what spatial computing is growing in its place, requires separating the concept from the hype cycle that distorted it.
This guide covers what the metaverse was and what it has become, what spatial computing is and how it differs from the original vision, where genuine traction is being found, and what the technology will realistically look like for businesses and consumers over the next three years.
What the Metaverse Actually Was
The metaverse, as Zuckerberg and other proponents articulated it in 2021, was a persistent, interoperable, three-dimensional internet where users would spend significant portions of their working and social lives. The vision combined virtual reality hardware for full immersion, blockchain technology for digital ownership of virtual assets, avatars as persistent digital identities, and vast virtual worlds that would host the social and commercial interactions that currently happen on conventional screens. Each element had legitimate technical precedent and genuine research investment. The combination was extrapolated far ahead of what any of the components could deliver in the timeframe implied.
The problems that produced the collapse were fundamental rather than cosmetic. VR hardware remained expensive, heavy, and isolating in ways that prevented the casual social use the vision required. Most people do not want to spend hours immersed in a headset, and those who tried reported fatigue, discomfort, and the specific loneliness of being physically isolated from the room they were physically sitting in. Interoperability between platforms never materialised, because each major player had stronger incentives to build a closed ecosystem than to contribute to a shared open infrastructure.
Digital assets backed by blockchain technology lost most of their value as speculative interest cooled. And AI, which offered immediate, practical productivity improvements without requiring new hardware or new behaviour patterns, absorbed the attention and investment that the metaverse had been counting on.
Spatial Computing: The Rebrand That Actually Means Something
The term that has replaced “metaverse” in serious technology discourse is spatial computing. Where the metaverse implied escape from physical reality into a digital one, spatial computing implies integration of digital information into physical reality. Where the metaverse required headsets, spatial computing works across a spectrum of interfaces from standard screens through augmented reality overlays to full immersion. The shift in framing reflects a real shift in what the technology is actually good at in 2026: augmenting physical work and physical spaces with digital information, rather than replacing them with virtual ones.
Meta’s own CTO acknowledged the stakes of this pivot directly. Andrew “Boz” Bosworth told TechCrunch that 2025 would be the year that determined whether Reality Labs achieved greatness or went down as a “legendary misadventure,” pointing to Meta’s Ray-Ban AI glasses, which had sold more than two million pairs, as the clearest sign that the pragmatic pivot was working where the immersive vision had not.
What Actually Has Traction in 2026
Apple has documented the enterprise pivot directly through its own developer platform, highlighting JigSpace and PTC as examples of businesses building genuinely useful spatial computing tools on Vision Pro. JigSpace, an enterprise spatial computing platform, reported roughly 50,000 Apple Vision Pro installations and 80,000 hours of usage by mid-2025, with the platform reducing sales cycles for immersive product demonstrations from six months to six weeks. That kind of concrete, measurable ROI is precisely what consumer metaverse platforms could never demonstrate and what enterprise spatial computing is beginning to accumulate.
Nvidia Omniverse, which provides industrial simulation and digital twin capabilities for manufacturing, engineering, and robotics development, has become one of the most practically significant spatial computing platforms without appearing in most metaverse coverage. Siemens, Schneider Electric, and ABB have all built significant industrial applications on top of digital twin technology that have concrete engineering and operational applications rather than consumer social ones. The common thread across all the cases with genuine traction is specificity: they address defined problems in defined industries with measurable outputs, rather than seeking to recreate human social interaction in digital form.
Generative AI has added a new dimension to spatial computing adoption. Users can now describe a virtual scene or object and have AI automatically generate the 3D assets for it, removing the technical barrier of 3D modelling expertise that previously limited who could create content for spatial computing environments. As LiveAIWire’s coverage of ambient AI’s real-world cost found, spatial computing is becoming a convergence point rather than a standalone technology: as AI becomes embedded in physical environments, the spatial layer for interacting with that embedded intelligence becomes more rather than less important.
The Hardware Reality
AR and VR hardware shipments declined approximately 12 percent in 2025 due to delayed product launches from major manufacturers, but analysts project approximately 87 percent growth in 2026 as new product cycles resume and accumulated demand is released. The longer-term trajectory shows a 38.6 percent compound annual growth rate for units shipped between 2025 and 2029. These numbers reflect a market that contracted through a product cycle trough rather than one that collapsed due to fundamental rejection of the technology.
AI-powered smart glasses represent the most likely near-term mass consumer entry point for spatial computing. Google’s partnership with Warby Parker for AI-powered smart glasses and Meta’s Ray-Ban AI glasses, which have sold significantly better than any of the company’s VR hardware, both point toward wearables that add ambient AI capability to physical reality without requiring users to adopt a new relationship with their physical environment. The glasses form factor has none of the social awkwardness and physical isolation problems that plagued VR headsets, which makes its adoption curve look more like the smartphone than the Segway.
What This Means for Businesses and Users
For businesses evaluating spatial computing investment in 2026, the relevant question is whether there is a specific workflow or customer experience problem where overlaying digital information on physical reality produces measurable value. If the answer is yes, the technology exists to address it and the cost of experimentation has fallen significantly. If the answer is “it seems like it might be useful eventually,” the track record of the metaverse era is the appropriate caution: eventual utility and current adoption are different things, and the cost of maintaining a virtual presence that nobody visits is the specific lesson of 2022-2024 that businesses should not need to learn twice.
The privacy implications of spatial computing, which requires sensing physical environments to overlay digital content on them, are covered from a rights and accountability perspective in LiveAIWire’s report on the digital resistance movement’s response to AI surveillance. Understanding how to design AI experiences that people actually keep using is directly applicable to spatial computing UX, which faces the same challenge as every previous new computing paradigm: the technology that works in a demonstration needs a different design approach than the technology that gets used every day. The metaverse failed the everyday test. Spatial computing is being built to pass it.
The Three-Year Outlook: What Is Actually Coming
The spatial computing landscape over the next three years will be shaped by three convergences. First, AI-generated 3D content will eliminate the specialist design bottleneck that has historically made spatial computing expensive to populate with content. Second, form factor improvements in smart glasses will bring ambient spatial computing to a mass consumer audience that has never adopted VR headsets. Third, enterprise digital twin adoption is expanding from manufacturing into urban planning, infrastructure management, and logistics, creating large institutional deployments that will drive the platform and tool maturation that consumer applications will subsequently benefit from.
The metaverse narrative failed because it extrapolated from what technology enthusiasts found exciting rather than from what ordinary people and businesses would actually use. The spatial computing narrative is more grounded, but it is not immune to the same failure mode. The applications that will define the next decade of immersive technology are more likely to look like digital overlays that make existing physical work easier than like immersive virtual environments that replace physical experience. That is less cinematically compelling but more commercially durable.
For individuals choosing which AI tools to invest in now, comparing the leading AI platforms includes understanding which are developing the most capable spatial content generation features, since that capability will determine which platform best supports spatial computing workflows over the next two to three years. The brands that integrate spatial content generation early will have the library advantage that determines which platform enterprises build on when the deployment environment matures.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity, and the social impact of emerging technology. He publishes daily at LiveAIWire.com.