Category report — 2026
Glasses that listen, translate, capture and project a small amount of information into the corner of your vision have moved from prototype halls to retail shelves. This is what actually shipped, what it does, and how the three leading approaches differ.
The shift
Heads-up displays, bone-conduction audio and on-face cameras have all existed for a decade. What changed in 2026 is the intersection: frames that weigh what glasses weigh, entry pricing that no longer reads as an experiment, and an assistant worth talking to.
What the hardware does
Microphones and open-ear speakers put a language model into the conversation you are already having. Live translation, hands-free teleprompting, and voice capture that returns a summary instead of a recording. On the newest frames the underlying model is selectable — different assistants for different tasks.
A waveguide places a small dashboard in the upper field of view: schedule, weather, navigation cues, notifications. High brightness and near-full lens transparency are what let it survive daylight without turning the world grey.
Footage from where you were actually looking — hands free, no framing ritual. It is also the most socially loaded feature on the device, which is why some makers now ship deliberately camera-free frames and others fit a hardware recording indicator.
Three approaches
These are not three versions of one product. Each maker resolved the same trade-off — how much display can you carry before it stops being eyewear — in a different place.
| Axis | Ray-Ban MetaMeta × EssilorLuxottica | RayNeo iORayNeo | XREAL AuraXREAL × Android XR |
|---|---|---|---|
| The bet | Eyewear firstNo display at all — audio, capture and an assistant inside a frame people already wanted to wear. | Glance, don't watchA minimal heads-up dashboard in a frame light enough for all-day wear. | Carry your screensA wide see-through display for real spatial computing, with the computer moved off your face. |
| Display | — none — | Waveguide heads-up display, high brightness, near-transparent lens | 70° optical see-through, micro-OLED per eye, electrochromic dimming |
| Camera | Yes — point-of-view capture with a recording indicator | Deliberately camera-free | World-facing tracking cameras plus a capture camera with a privacy LED |
| Where compute lives | In the frame, paired to a phone | In the frame, paired to a phone | In a tethered pocket puck running Android XR |
| Assistant | Meta AI, hands-free and multimodal | Selectable — several supported large language models | Gemini, with the wider Android app catalogue behind it |
| Worn | All day, indistinguishable from normal frames | All day — 33 g, ordinary-eyewear weight | In sessions — under 95 g and tethered, closer to a portable workstation |
Who wears them
Field technicians, warehouse staff, surgeons in training, cooks. Instructions and checklists in the corner of vision beat a screen you cannot touch with the gloves on.
Live translation is the first capability that non-technical buyers ask for by name. It works because it is passive: no app, no handing your phone to a stranger.
Hands-free teleprompting, and voice capture that returns a summary rather than an hour of audio nobody will replay.
The see-through class turns a train seat into a multi-window workspace, with the compute in a pocket instead of on the bridge of your nose.
Live captioning, navigation prompts and read-aloud are the same features as above, aimed at a use case where they stop being convenient and start being necessary.
Where it goes
The frames work, the displays are bright enough, the weight is solved. What remains unsettled is social: what is acceptable to record, what an assistant is allowed to hear, and whether a small amount of information in your vision makes you more present or less.