Qualcomm has revealed its latest AR glasses reference design, which it offers up to other companies as a blueprint for building their own AR devices. The reference design, which gives us a strong hint at the specs and capabilities of upcoming products, continues to lean on a smartphone to do the heavy compute, but this time is based on a wireless design.

Qualcomm’s prior AR glasses reference design was based on the Snapdragon XR1 chip and called for a wired connection between a smartphone and the glasses, allowing the system to split rendering tasks between the two devices.

Now the company’s latest design, based on Snapdragon XR2, takes the wire out of the equation. But instead of going fully standalone, the new reference design continues to rely on the smartphone to handle most of the heavy rendering, but now does so over a wireless connection between the devices.

Image courtesy Qualcomm

In addition to Snapdragon XR2, the AR glasses include Qualcomm’s FastConnect 6900 chip which equips it with Wi-Fi 6E and Bluetooth 5.3. The company says the chip is designed for “ultra-low latency,” and manages less than 3ms of latency between the headset and the smartphone. The company has also announced XR-specific software for controlling its FastConnect 6900, allowing device makers to tune the wireless traffic between the devices to prioritize the most time-critical data in order to reduce instances of lag or jitter due to wireless interference.

Though a connected smartphone seems like the most obvious use-case, Qualcomm also says the glasses could just as well be paired to a Windows PC or “processing puck.”

Beyond the extra wireless tech, the company says the latest design is 40% thinner than its previous reference design. The latest version has a 1,920 × 1,080 (2MP) per-eye resolution at 90Hz. The microdisplays include a ‘no-motion-blur’ feature—which sounds like a low persistence mode designed to prevent blurring of the image during head movement. A pair of monochrome cameras are used for 6DOF tracking and an RGB camera for video or photo capture. The company didn’t mention the device’s field-of-view, so it’s unlikely to be any larger than the prior reference design at 45° diagonal.

SEE ALSO
Meta Reveals VR Headset Prototypes Designed to Make VR 'Indistinguishable From Reality'

Like its many prior reference designs, Qualcomm isn’t actually going to make and sell the AR glasses. Instead, it offers up the design and underlying technology for other companies to use as a blueprint to build their own devices (hopefully using Qualcomm’s chips!). Companies that build on Qualcomm’s blueprint usually introduce their own industrial design and custom software offering; some even customize the hardware itself, like using different displays or optics.

That makes this AR glasses reference design a pretty good snapshot of the current state of AR glasses that can be mass produced, and a glimpse of what some companies will be offering in the near future.

Qualcomm says its latest AR glasses reference design is “available for select partners,” as of today, and plans to make it more widely available “in the coming months.”

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. See here for more information.


  • That’s amazing news…. without the cable these glasses will be much more usable!

    • silvaring

      True. I just wish AR glasses would finally have their DK1 moment !!! its been such a long wait haha

      • dk

        project nazare could become cv1

        • ZeePee

          Project Nazare is definitely CV1.

          Looking at the colurrent state of the tech, it’s looking very good for project Nazare in 2024.

      • ZeePee

        Nreal Light is already a DK1, almost DK2 level.

        Same with Magic Leap 2. That’s DK2 level.

        Project Nazare is CV1.

        • silvaring

          So whats the ballpark price for the CV1, and I expect it will be for a high end seated experience is that correct?

          • ZeePee

            It will have “experimental” free roam tracking ;)

            Price – wise my actual guess is nice large range of £800-£1500.

        • kontis

          No it isn’t. Not even close.
          What did DK1 did to HMD industry?

          Something very special. It offered absolutely horrible optical quality that was mocked at pro HMD convention as absolute useless trash by all the smug HMD veterans working for companies selling laughable amounts of HMDs.

          And then that “horrible” DK1, sold as a “dev kit” from one store, sold more units than any HMD in history.

          So why did that optically bad trash disrupted HMD industry? Because it made an absolutely magical trade off no one dared to do before: it gave up resolution, geometry and clarity for a gigantic increase in field of view. It was the first low cost HMD with FOV wide enough to do ORTHO STEREO. Nothing consumer-oriented could do that before in history.

          Analogically, the first AR DK1 moment should also be defined as:
          1. low cost – indie devs can afford it as an experimental gadget
          2. useful ortho stereo – “useful” is the crucial part. You can theoretically do it with any FOV, but anything below 80 deg is NOT useful for ortho. This was already explained as crucial point in one of the first presentations of Oculus for investors 9 years ago…

    • dk

      with the soc and battery the grasses get heavier and hotter and more expensive …..and at the same time it’s still sucking the phone battery

      splitting the tracking and rendering and whatever else is neat but a system like that still has problems …..and the problems would be less if there was a wire ……also after 3 hours of use or whatever u still have to plug in the glasses in an external battery

      • ZeePee

        Yeah it’s very cool, but dropping the wire before we have even reached CV1, is completely unnecessary.

  • sfmike

    I love VR and will be glad once developers put out and get over the fact that AR glasses aren’t going to be the billion dollar windfall the tech investment gurus have been telling them about for years within a couple quarters after release, we can all go back to developing better VR.

    • duked

      I think AR will succeed, but it will take 10-20 years before it’s as good as everyone envisions. And it will not replace VR, but maybe merge with it.

      • Definitely think it will merge, especially with the Cambria tech Meta is developing, if the cameras are good and high quality enough and color and have low enough latency I don’t know why you’d limit yourself to a device that can’t do VR in addition to AR other than for lightweight, stow-it-in your pocket use cases, which there aren’t many of

      • ZeePee

        Give it 5 years and we’ll have the AR we all envision.

        • kontis

          LOL. 5 years is NOTHING for this kind of R&D where you have to invent new things.

          In 2027 MR pass through will still be the best possible (but far too uncanny valley to be fun) “AR” you may get. Maybe with double the clarity of Cambria, and that’s it. So it will be like going form 720p to 1080p.

          I do remember all the expert predictions in 2015 and earlier about the VR systems we would have in 2022. Even the most pessimistic ones that everyone downvoted as concern trolling were BETTER than what is currently available.

          To be fair, some aspects are ahead of predictions (markerless tracking, standalone effective performance), but everyone expected raw specs to be leagues ahead.

          Jerri Ellsworth invented CastAR at Valve almost 10 years ago. Valve didn’t want it because it was too limiting for them (needs a table board).
          Currently (renamed to Tilt 5) those are STILL the most capable, widest FOV AR glasses that are possible to mass manufacture.

          All the more advanced hyped systems are still struggling after 10 years of hype with FOV and clarity that doesn’t even meet the minimum threshold of utility. And there is nothing indicating they can get to 90+ FOV in this decade either.

    • Ben Jacobs

      Not within a couple of Q’s, but they’re ultimately essential in regards to growing VR to a userbase comparable to console/mobile markets.

      It’s more or less the way to better VR, given how low the ROI currently is.

    • Tommy

      Personally, I have zero interest in AR and I think the market is a lot smaller than they think it is. Maybe that will change but I haven’t seen any must have uses for it yet.

    • kontis

      The hard AR concept absolutely could be gigantic and replace not just smartphones but tons of other things, like TVs, PC monitors and even movie theaters, but no one knows if technology good enough to make hard AR real will ever be invented.

      The Sword of Damocles HMD was invented more than half a century ago…

  • Cragheart

    I hoped for XR3.

    • Adrian Meredith

      It’s okay they rebranded their two year old processor to make it sound like a new one! F*CK qualcomm

    • kontis

      There is no XR3 planned. They changed their naming scheme. It’s more like Nvidia RTX x060 and x080. XR21 is low tier, XR2 is high tier. But they will have new generations of XR1 and XR2.

      The fact QUALCOMM is trying to BS so much with naming and marketing shows the marketers are probably now running the company, not the product people. Innovation is most likely in the back seat. And yes, I’m repeating after that famous Steve Jobs interview. It’s still very relevant.

  • ZeePee

    40° FoV is just a bit too small for anything like monitor replacement. Needs to be at least 50°.

    I think Meta will be aiming for at least 50°. In a form factor that’s as slim as this but even sleeker, much more like Nreal Air, and with a true Ray Ban design.

    • Ben Jacobs

      It’d be even more meta if you needed glasses to use the glasses ;)

    • kontis

      Normal people expect it to work like their normal prescription glasses with their full fov, so well above 100 deg. Anything less than that is a huge BUG and won’t be accepted. So definitely not happening in this decade.

      • XRC

        2030’s no doubt…