Independent tech analyst Ming-Chi Kuo reports that Apple’s XR and smart glasses roadmap will feature multiple XR devices, including a spec-bumped Vision Pro slated to release later this year. At the far end of the spectrum, Kuo also says Apple is making AR glasses, reportedly coming in 2028.

Kuo is a long-time tech analyst and respected figure in Apple product leaks. In a new blogpost, Kuo has laid out a timeline for a number of Apple XR releases, ostensibly based on various supply chain leaks.

Kuo reports that a new Vision Pro featuring the company’s M5 chip is scheduled for mass production in Q3 2025, with 150,000–200,000 units expected to ship. This next Vision Pro is reportedly upgrading the chip from M2 to M5, but will otherwise retain the original specs.

Kuo maintains the iterative approach is based on Vision Pro’s current position as a niche product, as the company hopes to use the next version to maintain market presence, reduce component stock, and further refine XR applications.

Photo by Road to VR

Next, Kuo says Apple is preparing a much lighter headset, reportedly called ‘Vision Air’, set to release in Q3 2027. Vision Air is said to be dramatically lighter at over 40% less than the current Vision Pro, which weighs in at 21.2–22.9 ounces (600–650 g), excluding the battery.

Kuo maintains Vision Air will include plastic lenses, magnesium alloys, fewer sensors, a top-tier iPhone chip, and be priced significantly lower to appeal to broader users.

The true next-gen Vision device is said to be ‘Vision Pro 2’ though, which Kuo says could arrive in 2H 2028, replete with a full redesign, Mac-grade chip, reduced weight, and a lower price, signaling Apple’s shift to move away from niche XR products and move toward mainstream adoption.

SEE ALSO
Valve is Building Key XR Components, Suggesting 'Deckard' Standalone is Moving into Production

Meanwhile, Kuo says Apple is investing heavily in smart glasses.

A Ray-Ban-like, audio-first wearable is expected to arrive in Q2 2027, Kuo says, noting that the company is hoping to manufacture 3–5 million units—ostensibly a significant push towards making its first smart glasses a mainstream success.

Oakley Meta HSTN | Image courtesy Meta, EssilorLuxottica

Like Ray-Ban Meta, and the recently unveiled Oakley Meta HSTN, Apple’s smart glasses are said to have no display, rather offering audio playback, photo and video capture, an AI assistant, and both gesture and voice controls.

Arguably the biggest claim among Kuo’s timeline is the mention of Apple XR Glasses, which the analyst maintains will include a color display (LCoS + waveguide) and AI features, making it the first true AR glasses from the company. Kuo says Apple is targeting release in 2H 2028, with a lighter variant being developed in parallel.

SEE ALSO
VR Design Unpacked: The Secret to Beat Saber’s Fun Isn’t What You Think

This follows reports of Apple potentially scraping a more casual XR glasses-type viewer, which would be tethered to Apple devices and use birdbath optics. The device was originally planned for release in Q2 2026, but Kuo maintains it was paused in late 2024 due to insufficient differentiation, especially around weight.

Granted, Apple is one of the most opaque black boxes in tech for a reason. The company historically announces products on stage, which typically also comes with a price and release date attached. While Kuo has a fairly reliable track record of reporting insider knowledge of Vision Pro, we’re taking this data dump with an equally-sized grain of salt.

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.

Well before the first modern XR products hit the market, Scott recognized the potential of the technology and set out to understand and document its growth. He has been professionally reporting on the space for nearly a decade as Editor at Road to VR, authoring more than 4,000 articles on the topic. Scott brings that seasoned insight to his reporting from major industry events across the globe.
  • Nevets

    Low specced, high-priced and served with plenty of Kool Aid. Some premium iPhones can't even run AI due to Apple's stingy memory. But they're the best and most excellently optimised things ever. Except when it comes to future proofing. Give me Samsung and similar any day. We owe a huge debt to Apple as regards smartphone evolution, but they're like your dad on the sofa now, wearing a jumper and watching Bowls.

    • Christian Schildwaechter

      The minimum requirement for using Apple AI is an A17 Pro or A18, which double the NPU performance from ~16/17 TOPS on A15/A16 to 35 TOPS. All iPhones using the A17 Pro/A18 also come with 8GB of RAM compared to 6GB before, so you can run Apple AI on iPhone Pro/Plus from 2023 or all iPhones from 2024 or later. But it's not just the RAM that's the issue here.

      After years of annoyingly keeping the basic Mac configuration at 8GB, Apple now switched everything to at least 16GB esp. to be able to run Apple AI, so the stingy memory policy and obscene upgrade prices certainly play a part. But a lot of people using basic M1 MacBooks with 8GB RAM from late 2020 still see no reason to upgrade, because the speed is fine, and thanks to very fast SSDs and clever MacOS memory management, these 8GB will get you a lot further than on a Windows laptop. Just not for AI, which needs very fast access to gigabytes large models, with swapping not an option. So these users won't get Apple AI, which many consider a feature instead of a bug.

      And usually Apple devices get a lot longer support than anything Android, which may count for "future proofing". No doubt they are low specced, high-priced and served with plenty of Kool Aid. But as Apple controls the OS, the specs are typically well balanced for the majority of users. And even the most basic configuration will usually work just fine for at least half a decade with continued OS updates, though you might not get all of the latest features, and even longer security updates.

  • Christian Schildwaechter

    TL;DR: 2025 M5 Vision Pro with the same issues, but more speed, for the enterprise market; 2027 A21 Vision Air for consumers, addressing the comfort/price issues, but otherwise offering the same experience as the current M2 AVP.

    A Vision Air makes a lot of sense, as it addresses high weight and price, which are the main issues most reviewers had/have with the AVP. By now Apple has learned how people use AVP, and can try to achieve the same experience with less/newer hardware. The M- and A-Series chips use the same types of cores, just in different configurations, with A-chips designed for low power mobile. By 2027 the M2 will be five years old, so a high end A-SoC should be faster, allowing for the same experience as the M2 AVP, but with less power/money/weight. And by reducing the huge amount of sensors to what is essential, they may be able to drop the extra R1 signal processor that works in parallel to the M2 in AVP.

    A more difficult question is who the M5 update is for. It not a redesign, mostly a SoC replacement with maybe some minor changes, similar to Pico 4 Ultra compared to Pico 4. But all it effectively does is bump up the speed, and this is the one area where AVP hasn't been widely criticized. Price/weight/power consumption/comfort all stay the same and will only be addressed by the Vision Air. So from a consumer perspective, they might as well have continued to sell the M2 AVP until the Air arrives.

    My guess is that this upgrade is mostly about pushing AVP for enterprise use instead of the consumer market. With no apps demanding it, regular customers won't really benefit from the extra power, but a large company running their own custom software might. Earlier this months Apple showed a couple of new features of visionOS 26, one of which was co-located AR, allowing several people using AVP in the same physical space to work together on virtual objects. This was demonstrated with Dassault Systèmes, who are a heavy weight in the aerospace CAD and simulation market and offer a number of compute-heavy engineering solutions. That's the type of software companies buy expensive workstation hardware for, and they most certainly would love every bit of speed they can get, money/power/weight be damned.

    If this is what Apple is going for, the next Apple Vision most XR users will care about will arrive in 2027 and no longer bear the "Pro" label. And what is released later this year will take the Pro more serious and explicitly target the enterprise market. Apple has done silent updates before, upping the specs without an actual product instructions, just offering a short press release. And if the M5 AVP is indeed aimed at the professional market, "the company historically announces products on stage, which typically also comes with a price and release date attached" may not apply here. So maybe don't expect any big announcement for later this year, or (mis-)interpret a lack of such announcements as Apple pulling back from XR.

    • deckert

      Excellent analysis and summary. I do think there will be more than just an M2-M5 upgrade. There are simple things Apple can do around the headstrap design (too heavy on the bridge of your nose/face) and upgraded battery. I do agree it will have more enterprise features (and would expect some partnership announcements with the big MDM players. I also think they will introduce multi-user device sharing.

      • Christian Schildwaechter

        Technically visionOS is based on iPadOS, which has multi-user profile functionality, but ONLY for MDM users like schools or companies. So it would make sense to also introduce it for AVP as an enterprise feature, as it most likely existed in the OS from day one. Just not for regular consumer peasants.

        I'm somewhat pessimistic regarding other hardware changes though, mostly because Kuo said that "the company hopes to use the next version to maintain market presence, reduce component stock, and further refine XR applications." We knew right from the beginning that AVP production was limited by Sony microOLED availability to at most 450K units per year. But they sold even less, apparently a lot less than they expected, and some time ago told their manufacturers to stop producing new components, planing to sell off existing stock until a successor arrives.

        So they may be sitting on a large pile of sensors, displays, cases and head straps not yet used, making them reluctant to offer some easy to achieve improvements. The estimated 150K-200K (per year/total?) M5 AVP units may all use mostly components they already ordered for the original AVP, with only the SoC being replaced. They could have stockpiled a few hundred thousands of these too, but AVP is now the last Apple device sold still using M2, everything else has been moved to a newer generation, which may be another reason for the upgrade besides more power for enterprise users.

  • xyzs

    Ok, if the guy is right that’s getting ridiculous…

    …3 more years to have a true successor to AVP.
    It better go directly 8k per eye, holographic lenses with varifocal and weigh less than 150g by then because that’s next century in nerd years.

    And 40 percent lighter for a vision air with plastic lenses in 2 years, that’s pathetic both in terms of specs (still way too heavy) and way too late..

    Apple is really a company falling apart.

    There is no excitement with them, since cook took over, it’s as dull as an insurance company.

    • STL

      Yes, Tim Cook is for Apple what Bob Iger is for Disney. Profit and wokeness, but no soul.

    • Christian Schildwaechter

      So for the 2027 Quest 4, we hope for 2.5K/3.5K microOLED, better passthrough, eye tracking, and XR2 Gen 3 with ~2.8x the graphics performance of the 2023 Quest 3, at roughly the same size/weight.

      And if Apple, who's 2024 AVP already offers 3.5K microOLEDs, excellent passthrough, eye tracking and ~50% more raw GPU performance than Quest 3, doesn't manage to one year after Quest 4 release an AVP2 that offers tech that doesn't even work as prototypes yet, at less than 1/3rd the weight of Quest 3, and technically completely unfeasible for anybody else, they are falling apart, because they are as dull as an insurance company.

  • STL

    In the last 20 years, Apple has invented nothing but iPhones.

    – Always faster iPhones, sometimes a bit bigger, sometimes a bit smaller
    – Very small iPhones with a strap (iWatch)
    – Very large iPhones (iPad)
    – iPhones you can strap to your face (Apple Vision Pro)
    – Large foldable iPhones with a keyboard (MacBook Air)

    Oh yes, and of course AirPods too!

    Innovation looks different than constantly just varying the same theme.

    • Now I Can See

      I'm far from a Meta fan, but I think Apple will be the Nokia to Meta's smartglasses. Apple's smartphone defined the last 20 years and overtook Nokia as the premier cellphone. It only took them about 4 years to knock Nokia off the "mobile throne". I'd say around 2029 we'll begin to see a shift away from the handheld smartphones and a shift towards handsfree AR glasses.

      • STL

        That's what I'm trying to say. Apple will not be able to do disruptive change.