Meta is retreating from the metaverse. VR teams are being cut, studios closed, budgets reallocated – from virtual worlds to AI and AI glasses. After 70 billion dollars in losses, it’s clear: VR hasn’t failed because the technology doesn’t work. It’s failed because nobody really wanted it.
It was meant to be the future. Mark Zuckerberg renamed his company, invested billions, bought studios, built hardware, dreamt of a world where people live, work and play in virtual spaces. The metaverse wasn’t just a product, it was a vision. And now, at the start of 2026, it’s history. Meta is cutting around 10 to 15 per cent of positions in Reality Labs, the division for VR/AR hardware and metaverse projects, in 2026. Affected are teams for Quest headsets, Horizon Worlds and several game studios like Twisted Pixel and Sanzaru Games. The fitness programme Supernatural is being closed. What remains runs on in “maintenance mode” – a polite term for “we’re giving up, but not switching off completely”.
Reality Labs has accumulated over 70 billion US dollars in losses since the end of 2020. In the third quarter of 2025 alone, the operating deficit stood at approximately 4.4 billion dollars against very modest revenue. In parallel, Meta is increasing investment in AI infrastructure and AI glasses – the Ray-Ban Meta Glasses. Marketing budgets are migrating from VR to AI glasses and generative AI. The message is clear: the metaverse was an expensive detour. Now it’s back to the main road.
The Numbers That Show the Failure
Worldwide VR headset sales fell by about 21 per cent year-on-year in 2025. For the full year, only around five million devices sold are expected, with a continuing downward trend at least into the third quarter of 2026. Meta remains market leader with Quest 3 and 3S, but shipments fell by around 16 per cent in the first three quarters of 2025 compared to the previous year – despite price cuts.
Revenue in the Quest Store is stagnating. The two billion US dollar mark was cited as early as 2023 and has practically not grown further by 2025. This points to absent content growth. Simultaneously, investment in AR/AI glasses and general AI is exploding. Funding for AR/VR start-ups collapsed by over 90 per cent between 2022 and 2023. The market speaks a clear language: VR is no longer a growth story.
Why VR Fails in the Mass Market
VR doesn’t fail technically. The headsets work, the graphics are impressive, the immersion is real. VR fails because people don’t want it. Or more precisely: because they don’t want it enough to overcome the hurdles.
First: low adoption and weak retention. Companies report low uptake rates. In the consumer sphere, many buyers hardly use their headset after the initial “wow phase”. It sits in the cupboard, gets charged too infrequently, is too cumbersome to put on. The barrier to entry is higher than with smartphone or console, and the added value doesn’t justify it.
Second: the content problem. “System sellers” with long-term motivation are missing. Many games are short, ports rather than native VR designs. Social experiences like Horizon Worlds never got beyond a small core community. There’s no killer app, no reason to get into the headset every day. And without daily use, no ecosystem gets built.
Third: hardware hurdles. Headsets are still bulky, cause motion sickness, sit badly with glasses and are impractical for everyday life. You can’t use them on the side, you have to sit down, put on, isolate yourself. That collides with everyday life – family, flat, office. VR isn’t “ambient”, it’s laborious.
Fourth: social and organisational barriers. VR is isolating. You can’t see what’s happening around you. You can’t simultaneously watch the child, answer the phone, open the door. In the enterprise environment, device management is complex, IT security complicated, training needs high. VR isn’t simply “there”, it has to be orchestrated.
The Strategic Pivot: From Metaverse to AI
Meta is shifting its story from “metaverse” to “AI-first”. The focus lies on generative AI, agents, recommendation systems and AI assistants on its own platforms – Facebook, Instagram, WhatsApp. AI smartglasses are being positioned as a more everyday-friendly bridging technology: lighter, more socially acceptable, “always on” and directly coupled to advertising, commerce and social graph – unlike isolated VR glasses.
The financial logic is clear: VR/metaverse generates gigantic CAPEX and OPEX without a visible path to mass usage. AI investments promise direct leverage on advertising revenue, productivity and new AI services. Politically and communicatively, the term “metaverse” largely disappears from Meta’s narrative. Instead, terms like “AI”, “agents”, “smart glasses” and “ambient computing” dominate. The metaverse isn’t dead, it’s just no longer important.
Is VR Dead or Just Repositioned?
Analysts speak not of a complete end, but of a long “trough of disillusionment” phase: low adoption, but where VR is deployed – training, simulation, medicine – ROI and satisfaction are actually high. Forecasts shift VR into a role as specialist technology for gaming, training, simulation and certain B2B scenarios, whilst the broad consumer market leans more towards light AR/AI wearables and traditional screens.
For Meta this means: less “moonshot” metaverse, more gradual integration of 3D/immersion into existing apps – 3D avatars, virtual spaces – plus AI features. So metaverse as feature, not as standalone platform. That’s the difference between “we’re building a new world” and “we’re adding some nice 3D elements to our existing world”.
What Remains
VR hasn’t failed because the technology was poor. It’s failed because the vision didn’t fit reality. People don’t want to flee into virtual worlds, they want to augment their real lives. They don’t want to be isolated, they want to be connected. They don’t want laborious, they want simple. And VR is many things, but not simple.
Meta has spent 70 billion dollars to learn this. Now it’s pivoting – to AI, to AI glasses, to technologies that fit into everyday life instead of replacing it. The metaverse was an expensive lesson. The lesson reads: the future doesn’t have to be spectacular. It just has to be useful. And VR was spectacular, but not useful enough.
When the future became a niche, the realisation remained: not every vision deserves billions. Some just deserve a place on the shelf.
