Android XR and the State of AR/VR/Spatial Computing

After teasing at various events that it was working with Qualcomm and Samsung on a VR headset, Google announced Android XR. Android XR is not a headset, but a broad platform for building eye-adjacent wearable devices with a variety of capabilities: everything from enhanced glasses to full VR headsets. Google is providing the operating system, APIs, and AI; it is not clear if Google’s Pixel hardware team will get involved at some point. Currently, Android XR is aimed at OEMs and software developers – there is no commercially available consumer or enterprise hardware available that runs on this platform.

That will change in 2025, as Samsung is planning to launch Project Moohan, a VR-with-passthrough headset that appears to be targeting the Apple Vision Pro. Early hardware has already been provided to developers (at least for demonstrations), and we should expect some teases about its capabilities and target audience at CES, MWC, or upcoming Samsung events. Samsung is not Google’s only hardware partner; Sony, Lynx, and XREAL are also listed on Google’s press release.

Marketing Challenges

Outside the industry, nobody uses, “XR” to describe virtual or augmented reality. The problem is that no term describes the range of experiences that Google hopes to enable with Android XR. Nobody knows how to describe these computers-for-your-face, especially when some of these things are more like earbuds than headsets. Apple calls Apple Vision Pro “spatial computing,” and that’s a reasonable description of an open-ended platform even if it hints that Apple isn’t sure what you’re going to do on it. Meta calls the Quest a mixed reality headset, but it is really an immersive gaming and exercise platform. Meta calls its Ray-Ban collaboration smart glasses, and that one is at least reasonable, but they are really camera and audio glasses. Meta has all but confirmed that it will be offering smart glasses with displays in 2025 but it isn’t clear if there will be a different marketing term for them.

“Augmented reality,” “mixed reality,” and “virtual reality” are all descriptive terms, but they can be limiting when applied to products that increasingly can do some or all of these. “XR,” or “eXtended Reality” acknowledges the continuum, at the cost of confusion. If Google and partners invest enough in clear marketing for products that use Android XR, the fuzziness of the name won’t matter. However, it does put more pressure on OEMs launching products to get that messaging right, because the platform itself doesn’t tell consumers or enterprise customers what your device can or can’t do.

Google’s History

Google is so notorious for launching and then abandoning products that there is a Google Graveyard website dedicated to chronicling them. Many of those killed products are predecessors to Android XR: Google Cardboard, Glass, North, Daydream, Tango, YouTube VR, Playground AR, and Soundstage. On the one hand, Google has certainly learned a lot about XR over the years – at least what not to do. But it is equally fair to ask if vendors should bet on Google sticking with Android XR long enough to matter.

Only time will tell, and there are certainly companies sitting this one out. However, if you do want to get into the market today, there aren’t many alternatives to Android XR. There are only a handful of platforms that have an established large application library that can be brought to bear and the resources to attract developer investment. Apple VisionOS is proprietary. Qualcomm’s Snapdragon Spaces is being folded into Android XR. Meta opened up its HorizonOS to OEMs, and Lenovo and ASUS have signed up. However, if you use HorizonOS, you will be definitely be competing directly with Meta’s own hardware, while a Pixel VR or Pixel Glasses are still somewhat hypothetical. Meta has been aggressive with pricing from the start, so the consumer market will be especially difficult to compete in, leaving ASUS to target gamers and Lenovo aiming at enterprise.

That leaves Android XR, or building your own OS. Samsung has tried turning Tizen and SmartThings into development platforms with mixed results – and failed with GearVR – but it has an enormous smartphone business built on Google’s platforms. For smaller companies like XREAL, Google is really the only answer; despite a market-leading position in display-enabled glasses, XREAL does not have the resources to build out a proper developer program. The only companies likely to roll their own APIs and potentially reach critical mass are Huawei and Xiaomi, and only for the Chinese market.

You could make the argument – best articulated by Stratechery’s Ben Thompson – that the future of XR is personalized, context-aware genAI applets for AR glasses. Maybe you don’t even need a library of 2D apps! However, that future is still in the future – genAI can’t understand what you see and react accordingly outside of scripted lab demos. (They are impressive demos.) Still, a limited, vertical approach to AR should lead to some interesting products in the short term – think the glasses equivalent of Pebble in early smartwatches. However, in the long term, it’s hard to bet against platforms that can attract third party apps and can leverage AI models developed by the largest companies in the world for the future. Android XR should be one of those.

Build Once, Deploy On Anything?

One key differentiator for Android XR is its breadth and flexibility. Google is enabling developers to use a single set of development tools and APIs to target everything from glasses without displays to 8K VR headsets. The computing can be done on device or split between the wearable and an external device (puck, phone, laptop, etc.). This sounds good on paper, and there are certainly advantages to standardizing around a set of development tools. However, I’m skeptical that a shared codebase will be at all practical. There are extreme differences in interfaces, computing capabilities, and even connectivity on different form factors. A directions app or a coupon app or a translation app are all going to need to be deeply customized to the device to be useful and efficient.

That said, the core idea that we will have different wearable XR form factors is sound. Things that look like glasses will be more limited in capabilities and immersion. They will likely be worn more often, for longer periods of time, and be more suited to outdoor use than headsets, no matter how good the passthrough on headsets gets. However, it is going to take a long time for microdisplays in glasses to match those in headsets, and power and battery life are more constrained. A headset is more likely to be worn for shorter periods of time, indoors, for richer computing or entertainment experiences. It’s not clear that a single software platform can address all the permutations, but there will be a continuum.

The original Samsung Gear VR, from my collection

Key Hardware Launch Partner: Samsung

Samsung is a technology innovator, and seeks inflection points to gain market position. It has done this at least twice effectively in its history, first using the transition from analog to digital TVs to vault forward in the home, and then with OS-based smartphones. Samsung can be patient: it tried every available OS until it hit upon Android and large screen sizes and grew to the global smartphone sales leader. Like Google, Samsung has a long history of trying and failing in VR, with multiple versions of Gear VR being the most prominent. That’s OK, Samsung will try again.

Project Moohan. This pic looks like a render, but there are real versions being given to developers.

The first Android XR device will launch in 2025 from Samsung. It’s dubbed Project Moohan (“infinity” in Korean), though it will probably be called Galaxy Glass or Galaxy XR or something less evocative at launch. This was a software platform launch, so details on the headset itself – or Samsung’s plans for other form factors – is not expected until later. Samsung did highlight some aspects of Project Moohan publicly. Its marketing materials reference both work and play; exploring, streaming, and planning in immersive environments; multimodal AI; gaze, gesture, and voice controls; “lightweight” and “ergonomically optimized; advanced displays with passthrough; and connected to other Galaxy ecosystem devices.

While this is obviously early, it appears that Samsung has been working on the hardware for a while: preproduction Project Moohan headsets were used at Google’s launch and have reportedly been given to select developers.

Key Silicon Launch Partner: Qualcomm

Android XR is very good for Qualcomm. Qualcomm has been investing in dedicated silicon for AR and VR longer than anyone. It powers the only real mainstream VR headset — Meta’s Quest — and the best-selling smart glasses — Ray-Ban Meta Smart Glasses. It is now a core part of Google’s Android XR across multiple form factors with a firm commitment for hardware from Samsung, one of the world’s largest consumer products companies. The only way Qualcomm loses is if Apple monopolizes the market to the exclusion of all others, or if AR and VR never take off with consumers beyond Meta’s success to date.

Qualcomm is being conservative in its estimates for XR growth; at its recent Investor Day it expected its XR revenues to grow to $2 billion by the end of the decade. I think that is unlikely: either XR will fail to meaningfully grow — there have been multiple stops and starts in the industry before — or it will be a lot larger than we $2 billion. I understand why Qualcomm doesn’t want to overpromise to investors, but put me on the optimistic end of the spectrum. Even if the broad consumer market doesn’t materialize, smart glasses are an obvious fit for many enterprise roles now that platforms are making it easier and more affordable to specialize. But I expect the consumer market to become real as well. We’re already seeing useful XR products on the market today, new AI models that can contextualize what you see are moving out of labs and onto devices, and Meta and Snap are showing off advanced AR glasses prototypes today that show what 2030 smart glasses could look like. Finally, Apple is “in,” lending urgency to other platform and AI owners to build out their own ecosystem before Apple iterates its way to dominance.

There are also clearly areas where Qualcomm – or rivals like MediaTek – have further opportunities in XR processors. MediaTek is already providing chips for Sony’s PSVR2 controllers, but the Android XR controller market is wide open. There is also clearly a need for low latency on-device image and spatial processing for smart glasses aside from Qualcomm’s Snapdragon AR2. Meta developed its own chip for its Orion prototype, and XREAL’s new One glasses use its own X1 processor to provide image anchoring features previously relegated to the original Beam external puck.

Techsponential is actively testing XR devices in-house from Apple, Meta, XREAL, Lenovo, and others, and we expect to get hands on with Android XR well ahead of launch. To discuss the implications of this report on your business, product, or investment strategies, contact Techsponential at avi@techsponential.com.