Updated: 2026-07-07

The extended reality hype cycle has been particularly dramatic in the last three years. Between 2023 and early 2024, with the announcement and launch of Apple Vision Pro, predictions pointed to a paradigm shift comparable to the smartphone. During 2025 the landing was hard: Vision Pro sold much less than expected, several flagship apps were pulled, Microsoft stopped making HoloLens 2, and Meta had to recalibrate Horizon Worlds forecasts several times, at one point nearly shutting it down in 2026. With the dust settled, time for an honest balance of what’s left standing and what’s been silently buried. One preview: applied AR, aimed at a concrete problem, is still the part that holds up best.

Key takeaways

  • Meta Quest 3S (around 75% of the consumer VR market, together with the rest of the Quest line) and Apple Vision Pro 2 (premium professional niche at 3,499 dollars) are the two poles of relevant hardware.

  • The real success story is in silent enterprise adoption: manufacturing, logistics, health, technical training, and remote maintenance.

  • VR gaming is reasonably healthy; the social metaverse is struggling, with Horizon Worlds cutting its in-headset creation tools during 2026.

  • WebXR has reached enough maturity for real cases without app download: AR catalogs from the browser, virtual tours, brief VR training.

  • Silently dead: HoloLens 2, Google Glass Enterprise Edition 2, direct sales of Magic Leap 2, metaverse as a unified concept, first-generation AR apps for commerce and advertising.

The hardware: lots of consolidation, little disruption

In device terms, the real picture is simpler than 2023 hype suggested. Meta Quest 3S (launched October 2024, 299 dollars) has become the dominant product in the consumer segment. Meta Quest 3, slightly higher tier, covers more demanding users and developers. Between the two, plus other Meta models, the brand controls around 75% of the VR headset market according to IDC, a figure that makes clear there isn’t really a "VR market": there’s a Meta Quest market with a few secondary players.

Apple Vision Pro remains alive but fitted in a very specific niche. After a launch with inflated expectations and disappointing sales, Apple recalibrated its strategy toward professional productivity and premium users. Vision Pro 2, unveiled on October 15, 2025 and on sale a week later, keeps the same 3,499-dollar price and adds the M5 chip, sharper displays and a 120 Hz refresh rate; it remains oriented toward professional uses where cost isn’t the limiting factor.

Chinese manufacturers, led by PICO and its PICO 5 alongside several smaller brands, have consolidated presence in Asia and Europe with technically competent products, though limited by regulatory restrictions and by the lack of a content ecosystem comparable to Meta’s. The second generation of lighter AR glasses (Xreal Light, Rokid, Lenovo ThinkReality) has matured as an independent category.

The enterprise story: where AR has stuck

The most interesting thing isn’t mass consumption but silent enterprise adoption. Industries like manufacturing, logistics, health, technical training, and remote maintenance have consolidated real use cases where AR and VR bring measurable value, often built on digital twins that model the factory floor before anyone touches it:

  • An AR-guided operator makes fewer errors in complex assemblies.

  • A remote technician resolves more incidents without travel, sharing video with 3D annotations.

  • A medical student practices more simulator hours without risk.

These economic returns justify investment in specialized hardware, custom software, and training, and have kept the professional industry alive while the general consumer lost interest.

Microsoft stopped making HoloLens 2 in 2025 (with software support guaranteed through 2027), and Magic Leap ended direct sales of its Magic Leap 2 on March 31, 2026, though authorized resellers can keep delivering units through the end of the year. Two of the names that carried enterprise AR have exited the hardware market almost at once, and the ecosystem still hasn’t collapsed: Lenovo ThinkReality keeps going in productivity, and several minor sector-specialized manufacturers fill the gap. Enterprise software support, with platforms like NVIDIA Omniverse and Unity Industrial Collection, remains an active area, though the underlying message is clear: dedicated AR hardware is a hard business even when the use case works.

The consumer story: games yes, social no

In consumer, the split is clear. The VR gaming segment is reasonably healthy, with successful titles like Batman: Arkham Shadow and sustained best-sellers. Small developers regularly ship titles that find their audience.

The social and metaverse segment, however, has clearly seen the sharpest regression in the whole sector. Horizon Worlds sums it up well: in March 2026 Meta announced it was shutting down the headset version and leaving it mobile-only, reversed course 48 hours later after creator backlash, and still went ahead on June 15, 2026 and disabled the in-headset creation tools, so anyone building or updating a world now has to do it from a phone. Behind the decision is a number that explains the urgency: Reality Labs, Meta’s division behind all this, has racked up more than 83 billion dollars in cumulative operating losses since 2020, 19.2 billion of that in 2025 alone.

Outside Horizon, the rest of the social segment doesn’t look much better. VRChat and Rec Room remain loyal niche communities but without explosive growth, and enterprise metaverses (Microsoft Mesh, Vision Pro workspaces) have had lukewarm acceptance. The promise of immersive sociability that dominated 2021 and 2022 has been replaced by a more modest reality of small communities around concrete experiences, and Meta itself is starting to talk more about Ray-Ban-style smart glasses than headsets as its bet for the future.

Multimedia consumption has gone better. Watching movies, recorded concerts, or sporting events in VR has a real market, though smaller than predicted.

WebXR: the open bet that’s still alive

One area deserving special attention is WebXR, the web standard for immersive experiences. By 2026 WebXR has reached enough maturity for real cases without app download:

  • Product catalogs in AR previewed from the mobile browser.

  • Immersive virtual tours without installation.

  • Brief VR training accessible from any compatible headset.

The technical ecosystem around WebXR, the W3C standard, has matured quite a bit: three.js with its XR module, Babylon.js, and A-Frame offer reusable component libraries that let you build decent experiences without a dedicated 3D team. Real portability between devices (Quest, Vision Pro with Safari, Android AR glasses) is comparable to traditional web: with careful testing and some tweaks, the same thing works on several devices.

A minimal example in A-Frame, the markup library for WebXR scenes, remains surprisingly short. This fragment builds a basic navigable VR scene from any headset with compatible browser. The technical barrier to start is genuinely low, and time from idea to functional prototype can be measured in hours, not weeks.

What’s silently died

Several recent bets have disappeared without noise:

  • Microsoft HoloLens 2: production discontinued in 2025, with software support through 2027.

  • Google Glass Enterprise Edition 2: sales stopped in March 2023 and support closed that September, much earlier than its late fame suggests.

  • Magic Leap One, and by 2026 Magic Leap 2 itself: the first was replaced by an enterprise model that has now also stopped direct sales, a sign of how hard the AR hardware business is even with real industrial clients.

  • First-generation AR apps promising to transform commerce and advertising, mostly pulled or reduced to occasional marketing experiments.

The metaverse as unified concept has lost almost all discursive relevance. Big strategic bets non-tech companies made in 2021-2022 (brands buying virtual land, immersive NFT galleries) have been closed or converted to something much more modest. The term is mentioned less and has been replaced by more specific language (industrial AR, VR gaming, multimedia immersion).

Conclusion

Extended reality is a perfect example of how hype distorts the valuation of a genuinely useful technology. XR isn’t going to transform consumer computing as the 2023 narrative promised; mobile remains the dominant device for accessing information and communication. But it’s not dead either: it has simply found the niches where it really brings value and is growing sustainably there, out of the media spotlight.

For a professional considering training or projects, the recommendation is specific by sector:

  • If you work in industries where AR has consolidated cases (manufacturing, health, maintenance, technical training), it pays to train in specific platforms and build capability; demand is real and growing.

  • If you work in web development and seek differentiation, WebXR is reasonable investment with moderate but consistent return.

  • If you think about mass consumption, keep distance: the market exists but is small, competitive, and not going to explode as predicted.

The general lesson is the same as with many other technologies: real value is found where concrete problems are solved better, not where hype paints the future.

This article is also available in Spanish: XR, AR y VR en 2026: el estado honesto tras el ciclo.