The Pimax Dream Air headset represents a new area of focus for the company. While most of its headsets up to this point have been necessarily bulky to achieve their signature large field-of-view, the Dream Air aims to make a headset that’s compact but still feature-rich. One of those features—a headstrap that automatically tightens—would be an industry first.

Priced at $1,900 and purportedly shipping in May 2025, Pimax’s Dream Air headset aims to take on an emerging segment of compact high-end PC VR headsets like Bigscreen Beyond and the Shiftall MaganeX Superlight.

Image courtesy Pimax

But it wouldn’t be Pimax if it didn’t make additional ambitious promises which risk pulling the company’s attention away from delivering its products on time and as promised. For the Dream Air, that additional promise is an optional compute puck which the headset can plug into to become a standalone VR headset. The company is calling the puck ‘Cobb’, and says it will include a Snapdragon XR2 chip and battery. Oh, and don’t forget the optional SteamVR Tracking faceplate.

Speaking of pulling the company’s attention… the announcement of the Dream Air continues Pimax’s trend of revealing new products before delivering on those it has previously announced. The company’s Crystal Super headset was announced back in April 2024 and originally planned for a Q4 2024 release, but is now said to be releasing sometime in Q1 2025.

As for the Dream Air, it will purportedly be compact and also full of a wishlist of specs and features:

  • Weight of 200g
  • Resolution: 13MP (3,840 × 3,552) micro-OLED per-eye @ 90Hz and “HDR”
  • 102° field-of-view
  • Inside-out tracking
  • Motion controllers & hand-tracking
  • On-board audio
  • Optional prescription lenses
  • Eye-tracking
  • Automatic IPD and automatic strap tightening
SEE ALSO
Samsung Reportedly Deepens XR Ties with Google in Push for Ray-Ban Smartglasses Competitor

That last one—automatic strap tightening—is a feature that hasn’t been included in any major headset to date. It’s an interesting idea considering the challenge of fitting a headset comfortably; many users want to crank their headset tight to their face so it won’t move, but the most comfortable way to use a headset is to balance tightness with stability.

The design of the auto-tightening strap also looks carefully considered. While we’ve only seen renders so far, it appears the tightening mechanism is hidden under fabric, making the tightening of the headstrap looks like it’s simply shrinking in place.

If the headset could effectively dial in the ideal tightness, it would be a boon for many users. Dream Air also has automatic IPD adjustment, which sets the distance between the lenses to match the user’s eye width (something most people also aren’t good at doing manually).

While it remains to be seen if Pimax can deliver something as svelte as promised, for now it looks like the company is flexing an industrial design muscle that’s been largely hidden by the utilitarian and boxy style of its previous headsets.

Image courtesy Pimax

However, Pimax isn’t giving up those boxy designs of yore. The company says that a compact headset is a new area of focus for the company, but it will continue developing its larger and wider field-of-view headsets.

Pimax is already taking pre-orders for the Dream Air, with a price of $1,900 and an expected release date of May 2025.

Newsletter graphic

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. More information.


Ben is the world's most senior professional analyst solely dedicated to the XR industry, having founded Road to VR in 2011—a year before the Oculus Kickstarter sparked a resurgence that led to the modern XR landscape. He has authored more than 3,000 articles chronicling the evolution of the XR industry over more than a decade. With that unique perspective, Ben has been consistently recognized as one of the most influential voices in XR, giving keynotes and joining panel and podcast discussions at key industry events. He is a self-described "journalist and analyst, not evangelist."
  • guest

    They are better known in China and for apps that require wide FOV

  • Christian Schildwaechter

    The resolution of 3840*3552 matches the MeganeX Superlight 8K (3552*3840, rotated by 90°?) Other display specs also match (90Hz, HDR/10bit), so both very likely use the same microOLED displays. Both HMDs are priced at USD 1900, but for that Pimax includes motion controllers, integrated audio, eye and inside-out tracking, all missing/adding extra cost on the MeganeX (with a flip-up display, adjustable diopter and eye relieve).

    For years the MeganeX with many prototypes and missed shipping dates seemed vaporware, but 2024 they both massively increased resolution and started taking (pre-)orders. Pimax still isn't a beacon of shipping reliability, but the Pimax Crystal came with all (?) the promised features incl. eye tracking and onboard XR2, and seems to be a great HMD.

    The probable reason is availability of standard components. In the past a new HMD kind of required starting from scratch. Now companies have access to 3rd party microOLED displays, pancake lenses, Android with OpenXR stacks, XR2 SoC also usable for streaming, lores nIR tracking cameras connecting to the XR2's CSI ports, and hand tracking software using them. So creating a new HMD has become much easier.

    This could mean that more technically similar HMDs will appear soon, maybe from startups that, following the AVP release hype, collected venture capital from investors suddenly wanting to jump into high end/price productivity XR. The effect could be reduced prices and headsets for more diverse use cases to set them apart. As usual, a market with actual competition should be good for users.

    • Dragon Marble

      Competition doesn't always benefit the users. For example, would you rather have Google Earth available on Quest and Vision Pro as well, or would you prefer Google keeping it for itself as a competitive edge? There are many other examples where the users would've benefited more if the companies had cooperated instead.

      As to these smaller headset makers, they are actually competing to make themselves better instead of sabotaging each other. Even in this space, however, I would rather have a big player consolidate the high-end PCVR market so that we can have proper support behind the headset in order to turn good ideas into a great product.

      Bigscreen, Pimax, MeganeX. There's something I really like about each. But most likely I won't buy any.

      • Stephen Bard

        Strange that you should mention Google Earth as an example of something with no competition since the EarthQuest Quest app is actually better than the PC Google Earth, with many more features, and both Wooorld and Fly are also quite good. How would you expect the various high-end headset makers to "cooperate"? I keep wondering why so many companies bother to build these very expensive micro-OLED headsets that have obsolete claustrophobic FOVs and marginal brightness. No matter how beautiful the resolution/color is, I would never consider buying a headset with FOVs less than the Quest 3, because even that feels barely adequate. Why can nobody make 4k micro-OLED displays that are 25% wider? And while I am complaining, why can't someone make a proper micro-OLED display for the many AR glasses all of which have an FOV of less than 50 degrees? (and resolution higher than 1080).

        • Bram

          Totally agree on the need for bigger fov! But unfortunately, mass production of micro-oled displays at affordable cost is hard! For production they use equipment that's normally used and build for the semi-conductor industry. The wafers used have a limited size and producing bigger micro-oleds to get bigger fov's comes with fast increasing risk for production errors, faulty pixels, color inconsistenties etc and thus increasing costs substantially, that are already high with the current sized micro-oleds. It made Meta stepping away from plans to bring out a next quest pro with micro-oled displays in 2027. It would become too expensive and likely have an fov smaller than the quest 3, which would be a step back. Another reason not to go for bigger fov's is that a lot of pixels of your display will be assigned to the periphery of your vision, lowering the pixeldensity in the centre.
          Still there are companies working on producing pancake lenses that can deliver a higher fov even with the small micro-oled displays we have now. Fingers crossed, there is still hope :))

          • Good analysis. Which companies in particular are working on high FOV pancake lens technologies? Got a link for further research or information?

        • Christian Schildwaechter

          TL;DR: High FoV microOLED displays aren't more difficult to produce than the currently available ones, as it only depends on the lenses, but with pancakes you have to choose between small HMD size and large FoV; companies go with 90°-100° because a higher FoV also requires a much higher rendering performance, which is sort of wasted on the least visible parts of the image when it is already challenging to render for a 4K display.

          You can absolutely create a microOLED display with high FoV, because the FoV doesn't depend on the display, but the lenses. And you can also use pancake lenses to achieve a larger FoV.

          Pancakes are basically just a set of consecutive lenses working together, causing less optical distortion than using just one lens with a more curved surface would, which is one of the limiting factors for achieving higher FoV. The ones in VR HMDs use internal reflection to significantly reduce the required distance between lens and displays, allowing for much slimmer headsets. You could use the set of lenses to achieve a higher, distortion free magnification instead, which would give you a larger FoV, but then you'd again need to make the lens-display distance larger, resulting in bulkier headsets.

          You can basically use pancakes to either reduce the headset size or increase the FoV, but not both at the same time. And so far manufacturers go for the smaller size, which makes kind of sense with high resolution displays. Making the FoV larger isn't free, it comes with exponentially increasing render cost for the least important parts of the image. Most of the time we look forward because that's where our eyes are the sharpest, while the peripheral vision is seen at a much lower resolution and only very muddy. If something interesting pops up there, we check by moving our eyes and then turning our head for the clearest image.

          Driving a 4K display already requires a lot of processing power even when limited to about 90°. Due to perspective distortion, adding a little bit of FoV requires rendering a lot more geometry in two screen directions that most users will never ever see, or only very fuzzy. Looking at a flat display with lenses causes extra distortion that make this even worse. So while we don't have clever ways to reduce the render load for the peripheral vision or start using dome shaped displays, it makes sense to throw all the compute power at the 90° that we can actually see somewhat sharp by moving our eyes.

          • Quest 3's 110 degress FOV is just about the sweet spot it would seem to me. Especially for when I'm using Meta's "Open Periphery" facial interface. It naturally opens up my FOV to the pass through which I'm using 99% of the time since I'm not gaming but rather using Quest for media consumption. I've also found that "The Spatial App" is my goto app for this use case.
            I'm sincerely hoping that Lenovo (or Asus perhaps) is able to deliver a horizon OS based high density micro-OLED based headset in 2025. They just need to take the Quest 3 as it is today, replace the XR2 with XR2+, replace the LCD displays with these BOE 4K per eye micro-OLED displays, add eye tracking for dynamic foveated rendering and upgrade the pass through camera resolution. Deliver that in virtually the same Quest 3 shell and I'm a happy camper for some time to come.

          • Christian Schildwaechter

            All the tech required for Asus/Lenovo to do that is already available, even the eye tracking and ETFR support in Horizon OS. It might get difficult to squeeze all that into the same Quest 3 shell though, mostly due to extra heat. If they move the battery to the back and make it larger, they may have the power budget for more active cooling.

            That would also help with running the XR2+ Gen 2, mostly an XR2 Gen 2 with 15% higher CPU and 20% higher GPU clock, as fast as possible without underclocking anything, which would be necessary to drive the almost four times the pixels a 4K display offers compared to Quest 3, even assuming (much) more efficient ETFR than on Quest Pro.

          • I'm 100% tethered already. I sit in an Eames lounger and have the headset plugged into its charger. The cable goes around the back of the recliner and I never notice it again after putting the headset on. I also have no range anxiety b/c I know I have continuous power. So for my purposes, they could completely shift the battery to an external puck. Maybe if they need to include a 1/4 size of the current quest inside the headset, just for standby power while connecting the puck or whatever. I'm fine with either.

            Would I like a sleek redesign similar to this Pimax render or the Google/Samsung "Moohan"? Yes, that would be great and that's what I'm expecting Lenovo or Asus to deliver. I'm just saying if they took Quest 3 shell and replaced the components one for one with the upgraded options, that would be my ideal headset today and some years into the future.

            I think the 3.5K micro-OLED display is the rubicon for a "virtually real" experience in VR. That's the standard these headsets will need to start from if they want to deliver on the Vision Pro experience and improve upon it.

      • Christian Schildwaechter

        Cooperation doesn't exclude competition or vice versa, like standards for exchangeable parts still allow outdoing others. Multiple competing headsets, connecting to lighthouse trackers/controllers and running Steam PCVR games, were only possible because companies didn't having to establish whole new eco system each time.

        Google Earth is more about abusiving exclusivity than cooperation bundling/saving resources. Google sells geo-data/navigation services, and Earth like apps on Quest pay for using their APIs and data. Company websites embedding Google maps also pay, but competing map services from Microsoft, Apple, HERE, Yandex, OpenStreetMap etc. ensure Google can't dictate prices or access.

        But only Google, Microsoft and Apple had the resources to create 3D views of (parts of) the world. MSFT is only possible because Microsoft doesn't have to pay Google for data like Quest apps. SoI'd rather have competing geo-spatial data services allowing to create mapping tools or flight simulators on multiple platforms, filling gaps and niches. Not one occasionally benevolent large player and then depend on it either releasing Earth on a platform because it serve its own interests (preventing competition), or keep it as an exclusive, or simply avoiding the cost.

    • optimisticguy

      So would you agree that this product will turn up in 2025? Reusing the software from crystal light and repurposing the micro oled from crystal super?

      I might be crazy but I think there's a good chance it might actually arrive before the immersed visor lol

      • Christian Schildwaechter

        I'd say it is possible, and increasing competition will make companies hasten to actually release products. MeganeX pre-orders were announced two months ago, during which it was in many ways the technically best PCVR HMD one could at least order. Now some potential buyers may already be hesitant and wait if the Dream Air actually ships.

        And the moment Samsung releases their XR2+ Gen 2 based standalone with similar resolution at a similar price point, everyone will be forced to react and might have drop the price for tethered PCVR headsets lacking their own processing. So being fast means making more money, which should provide some motivation for Pimax to get their act together and release an headset in 2025 as early as possible.

        • Jose Ferrer

          But most likely the Samsung XR device will be missing a DP wired connection, so even at the same resolution the image will need to go through the 200Mbps decoding limit of the XR2+Gen2 chip.
          In any case, as a PCVR player, I always prefer the VR manufacturers who separate the vision device (Air Dream, BigScreen, Maganex8K, Visor, PSVR2, Index, etc) from the computing thing (PC, laptop, Cobb, Console? mobile?, etc) so the vision device is lighter, cheaper and can be used wired for multiple purposes (PCVR, Android XR?, )

          • Christian Schildwaechter

            It's less that the Samsung XR HMD would be targeting the same user group, and more that Samsung and Pimax/Shiftall offering HMDs in a similar price range, with Samsung offering a full standalone device that obviously is more expensive to produce, will lower the perceived value of the tethered HMDs. That may not matter if your main concern is the best possible image and therefore DP-In, but first Quest 2 and now increasingly Quest 3 being the most popular PCVR HMDs shows that for the majority the perceived value is more important than a low latency, compression artifact free image.

  • Jaap Grolleman

    Hey man, we have loads of satisfied users across a huge range of headsets, across now 9 years. Our previous headsets are very popular among simmers and not always feature that prominently on VR media.

    • Hi Jaap, this is an ambitious and very interesting announcement. Can you comment on whether or not there exists a functional prototype and if so why it was not shown during this video announcement instead of a simple black 3D printed model?

    • Alrighty ….

      Thanks for responding.

    • Dodo Zapp

      Hi Jaap, I think you will have enough insight to appreciate that a product clearly still being in the design phase, will not be ready to ship in May 2025. It is not just a small spin-off a la Crystal Light or Artisan, it‘s a new design only re-using the optical MicroOLED module from the Super.

      You guys simply CANNOT be taking pre-orders with 1,100$ pre-payment now for a mere concept stage product! This is bound to turn into another fiasko like the trade-in induced purchases of „old stuff“ with the 12K being promised, which turned out to be vaporware and now customers are stuck on headsets they would not have bought otherwise.

      I owned a 5K+, 8K and still have an 8KX but I am sceptical if I see Pimax announce new headsets time and again while they have not even released their last announced headset.

      The result is a lack of focus. Pimax jump to the next headset instead of taking care properly of the toothing pains of their current headsets and polishing them, which shows in the reports on quality issues.
      Or they abandon them right away, like it happened to the og Crystal, whose stand-alone capabilities join Pimax Portal on the graveyard of ideas sold to customers for good money before pulling the plug quickly.
      Even if the Dream Air materializes (realistically in 2026), will it have proper QA? Will most of the R&D support internally have been redirected already to the next 1-2 successors/new products already being jumped on.
      I don‘t see enough focus & continuity to make me want to spend 2k with taxes on a headset from Pimax.

  • Michael Speth

    "Auto-tighting" is a disaster waiting to happen. How many people are going to get injured before this product is pulled?

    PSVR2 implemented the ideal strap solution including the best way to tighten it – using a wheel. Pixmax can't even get their current projects correct let alone this feature which is a death trap waiting to happen.

    • The design Apple patented is just about perfect and they opted for using a manual tensioner – it works extremly well and I've never thought it needed to be powered or automatic. I like it so much I made 3d printed adapters to use it on my Quest 3.

      Pimax appears to be seeking to one-up the Apple design here but its unnecessary and overkill. No one is really asking for this, its gold plating in the extreme. Everything else about this product is spot-on but for this and the over promise of a May release without even showing a prototype at the announcement.

      • Michael Speth

        The apple head strap is garbage, both of them.

    • Dale Kirkley

      "Death trap".

      Looks like we have someone who doesn't understand torque, and how much is required to cause bodily harm.

    • Nevets

      My God, you're right. This technological garrotte is going to trepan dozens of unwitting users AND BURST THEIR SKULLS before the authorities realise what the hell is going on and finally pull it from the market, ban it forever and indict Pimax on charges of corporate and actual manslaughter. We have to act now before it's too late. #nervegear #sao #palmerluckey

  • We all felt the need of the announcement of ANOTHER Pimax headset lol. It seems they can't live without announcing a new device every two months :). Anyway, I'm curious about this headset, I would like to try the auto-tightening feature.

  • patroza

    Please find a way to connect this to iPhone 15+, iPad Pro and other usb-c/dp compatible phones! Even if it’ll be just a 2D screen. Even though the built in X1 chip in XReal one is pretty cool, all I really want is unplug headset from pc, plug into phone, and continue where I left off on Apple TV etc.
    The phone or tablet is the device you already carry with you. Just like the phone took over most of the camera function.

    • Christian Schildwaechter

      The problem is that the mobile XR2 SoC don't come with any decoders that could receive an HDMI or DP or USB-C DP-Alt mode signal and simply pass it to the GPU to be displayed either directly or projected inside XR.

      So you need either custom software on both phone and HMD, or a license to receive the proprietary Apple or Google casting protocols, or use reverse-engineered that is so far incomplete with compatibility issues, when you only want to use software and connect wirelessly. Or you add a separate decoder/path to the frame buffer like the Pico 3 Neo Link with a separate USB-C connection for DP in did. Or you integrate a tiny HDMI grabber similar to what is now possible with Quest 3 that can display an MP4 stream coming from a grabber connected to it's USB-C port as a separate window inside the headset.

      While all of that adds effort and/or cost, it is certainly possible. And with several headsets now competing in the higher end PCVR market with prices close to USD 2000 that should come with a healthy margin, there is a need to provide unique selling points. Being able to integrate phone content would be a pretty good one, so there definitely is hope that some of it will happen.

      • patroza

        But I’m talking about the Dream directly.
        I mean if I need their XR2 device too, i would buy it. But it’s a DisplayPort usbc device so it should be able to direct connect to my phone, just lik XReal/Viture/etc glasses can

        • Christian Schildwaechter

          TL;DR: unprocessed DP-in works for low FoV displays without a lot of tracking capabilities, but a VR HMD like the Dream Air will today come with some onboard processing for handling sensors and maybe even reprojecting a flat image coming from a phone for a proper display in VR.

          The Rift S and WMR headsets got away with the PC doing all the tracking, connection several lores nIR cameras via USB through a rather thick and expensive cable. That gets less and less an option with more sensors on the headsets and higher resolution cameras for passthrough etc., and everyone trying to use standard (cheap) USB-C cables for transferring both video to and sensor data from the headset.

          So nowadays you typically have some SoC onboard to both handle sensors and connect to the controllers as well as doing some tracking. The PSVR2 connects via USB-C with DP-Alt mode, contains a SoC that's smaller than the XR2 Gen 1 in Quest 2 and apparently manufactured by MediaTek for Sony. It communicates with controllers and haptics, apparently does (parts of) the controller, eye and now probably also hand tracking. It might integrate Tobii's EyeChip that's otherwise missing on the PSVR2 PCB, and also decodes the DP signal coming over two USB-C data lines. It also integrates a GPU and is capable of showing a video signal send from a regular source as a kind of virtual cinema screen, correctly projected inside VR with head tracking by itself, even when not connected to a PS5 or a PC with Sony's software.

          Similarly the Dream Air will have to include some type of SoC that allows to connect several room/controller and eye tracking cameras, preprocess the data and send it to the PC, unless they go with a very thick cable or some solution like the Pimax fibre optics DP cable with merging lots of data lines at very high speed. DP-in alone isn't enough. And if you need compute on the HMD anyway, it makes most sense for it to also receive the DP signal and display it, like the PSVR2 does.

          But creating a custom SoC isn't exactly cheap, so headsets selling in low numbers will very likely go with one of Qualcomm's existing XR chips instead, which unfortunately all lack the option to receive DP and pass it to their integrated GPU. Connecting a phone directly works on the XReal etc. because these are basically just low FoV displays without lenses that would adding geometric distortion, most of them including only 3DoF tracking from a IMU or no tracking at all, and none doing any type of internal reprojection. The Dream Air offering inside-out and eye tracking means there are at least six additional cameras involved. And the use of lenses and a larger FoV requires that any rectangular view coming from a phone will have to be re-projected by a GPU to again look flat inside the HMD, a task that for Rift S or the Pico 3 Neo Link with DP-in the connected PC was expected to do before sending the digital video signal to the HMD.

          • patroza

            Thanks for comprehensive response.

            I mean you can disable all the tracking and leave just the displays.
            In terms of lenses; the Goovis g3 max has lenses, while taking usb-c dp input up to 1440p60 or 1080p120, on micro oled at 65 degree fov; 15+ more than XR glasses and 40 less than common VR glasses that’s true.

            As to good usb-c cables; they can be very thin. Optical especially, or short distance copper.

  • Adrian Meredith

    The pimax Vishun pro

  • Anonymous

    I don't understand the appeal of standalone if they aren't willing to spend on funding (devs), expanding and maintaining a store, or at least team up with another bigger player like Horizon / Android XR / Pico store?

    • Dodo Zapp

      The hidden value is that you have an untethered headset for PCVR by using Virtual Desktop or Steamlink.

      • Christian Schildwaechter

        The standalone unit is also optional, so the Dream Air clearly remains a PCVR HMD intended to run Steam games. Only with a wireless add-on that in theory can also run standalone apps, but is rather underpowered for 4K, with no existing apps. Pimax might add some themselves, like a virtual cinema mode not depending on being close or connected to a PC, which would be actually useful as a mobile application. The compute unit may also be interesting for some special use cases and custom developed standalone apps, but doesn't look like Pimax attempt to compete with Meta in the consumer stand-alone market.

        So going through the effort of creating a store and then investing lots of money to get dev support for the minuscule user base that would even want to run stand-alone games on a 4K PCVR HMD, really makes no sense. Those with matching use cases will be happy if the option to create custom apps even exists, but the concept seems very different from for example Meta's Orion where the separate compute unit actually provides the main functionality due to the extremely restricted power and features available on the headset itself.

  • FalconLX 911

    I have been scrouged over one too many times by Pimax- I will wait until it is available on Amazon or a reputable vendor that offers no worry returns. Unfortunately by the time that happens better main stream headsets will be available.

  • xyzs

    This is just fiction. They don’t even have a prototype.
    So basically it’s just their Christmas list of a headset…

    This company su@ks, they release products after products without doing things proper, and while they have a pending product, they announce the next one.
    They claim they can score top specs but they can barely make a decent website and promotional videos.

  • Nevets

    There's something about Pimax hardware announcements that makes me never want to read beyond the headline. I can't put my finger on it. I LOVE hardware announcements.