Steam Frame isn’t Launching Alongside Steam Controller on May 4th Due to RAM Shortage

40

Steam Controller, Valve’s next-gen gamepad, is slated to launch on May 4th for $100, although you shouldn’t expect to see a big ‘buy’ button next to Steam Frame or Steam Machine.

Speaking to Polygon, Valve revealed that it’s only releasing Steam Controller next month for a pretty important (and slightly obvious) reason: Steam Controller “doesn’t have RAM in it,” Valve hardware engineer Steve Cardinali told Polygon.

“We wanted to build up quantity so that we could try to address everybody who wants one at launch,” Cardinali maintains.

Image courtesy Valve

Notably, Steam Machine is set to include at least two bundling options: one with a Steam Controller and one without, which could put a kink in Valve’s supply efforts to produce enough Steam Controllers, as Machine specifically features built-in support for the gamepad in an effort to make it more of a living room console.

Image courtesy Valve

In an IGN interview, Valve’s Pierre-Loup Griffais spoke circumspect about why the company isn’t pushing out all three products as previously planned.

“For us, the controller is something that stands out on its own and we want to make sure that we can get that to customers in parallel to anything that might be happening with Steam Machine.”

SEE ALSO
Samsung Reportedly to Debut First Smart Glasses at Galaxy Unpacked on July 22nd

While Griffais doesn’t specifically talk about issues with component sourcing, which have spurred RAM and storage prices to exponentially increase over the past year, Valve confirmed as much in February, noting the company had to “revisit” the pricing and release dates of both Steam Frame and Machine.

That said, Giffais again echoed that Valve doesn’t have exact details about the timeline (or price) for Steam Frame or Steam Machine, although he says the company is “hard at work on trying to get them out the door. I think we are definitely expecting to roll out some news soon about that, but in general, I think things are going well.”

Looking for more Steam Frame news?

Valve Unveils Steam Frame VR headset to Make Your Entire Steam Library Portable: Valve shows off Steam Frame, the standalone headset that can stream and natively play your entire Steam library—with only a few caveats right now.

Hands-on: Steam Frame Reveals Valve’s Modern Vision for VR and Growing Hardware Ambitions: We go hands-on with Valve’s latest and greatest VR headset yet.

Valve Says No New First-party VR Game is in Development: Valve launched Half-Life: Alyx (2020) a few months after releasing Index, but no such luck for first-party content on Steam Frame.

Valve is Open to Bringing SteamOS to Third-party VR Headsets: Steam Frame is the first VR headset to run SteamOS, but it may not be the last.

Valve Plans to Offer Steam Frame Dev Kits to VR Developers: Steam Frame isn’t here yet; Valve says it needs more time with developers first so they can optimize their PC VR games.

Valve Announces SteamOS Console and New Steam Controller, Designed with Steam Frame Headset in Mind: Find out why Valve’s new SteamOS-running Console and controller will work seamlessly with Steam Frame.

Steam Frame vs. Quest 3 Specs: Better Streaming, Power & Hackability: Quest 3 can do a lot, but can it go toe-to-toe with Steam Frame?

Steam Frame vs. Valve Index Specs: Wireless VR Gameplay That’s Generations Ahead : Valve Index used to be the go-to PC VR headset, but the times have changed.

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.

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.
  • deHavilland

    I assume in the sentence 'Notably, Steam Machine is set to include at least two bundling options' you mean 'Steam FRAME'?

    • Andrew Jakobs

      No, Steam Machine.

    • ZarathustraDK

      Steam Machine makes sense. Options would be: Steam Machine + controller, and Steam Machine + Frame.

      • hamburgargurka

        I would argue that the Frame is not an amazing candidate for the Machine because of the low 8GB of VRAM. It won't deliver a smooth experience for the most popular VR game that is VRChat, and that isn't a good look for Valve.

        • Herbert Werters

          For which VR games and which 2K VR headset is 8 GB of VRAM not enough? Could you explain that to me?

          • hamburgargurka

            Like I wrote, VRChat easily demands more than 8GB on lots of worlds, especially if you want to avoid aliasing with MSAA or supersampling.

          • NL_VR

            VRChat is probably not a priority in that case.

          • hamburgargurka

            It's the most popular VR game, and VRAM requirements will only increase as time goes on. VR uses more VRAM than traditional gaming, there's no way around that, and you're already seeing certain new PC games struggling with just 8GB of VRAM. 8GB rigs shouldn't be normalized or excused today, sorry.

          • NL_VR

            The the sollution is to lower resolution and graphics settings. its PC we are playing on after all. 1 year ago i used a 8GB card

          • Herbert Werters

            For which VR game"S"?
            Exceptions prove the rule.

          • hamburgargurka

            I only mentioned VRChat, you're the one who brought up plural. Either way VRAM requirements are only increasing for modern titles, and a new console generation is around the corner sometime next year which will further put strain on 8GB GPUs.

          • Herbert Werters

            Yes, but you’re taking a single game and generalising it to everything else, claiming that it doesn’t reflect well on Valve simply because one very specific game with very specific, overloaded maps might be problematic. How else are we supposed to interpret a statement like that? Explain that to me.

        • Christian Schildwaechter

          TL;DR: you forgot to consider how much of a resource hog Windows is compared to Linux.

          Just a few days ago, Valve engineer Natalie Vlock released a Linux kernel patch that optimizes VRAM usage by prioritizing the current foreground process/game, preventing background processes from allocating VRAM, and thus keeping the game fully in the fast GPU RAM, preventing expensive swaps to system RAM. It's basically improving VRAM management, which doesn't matter if you have a lot of it, but can have significant impact on systems with only 4GB or 8GB of VRAM.

          This is very obviously an optimization designed esp. to improve the SteamOS performance on the Steam Machine with only 8GB of VRAM. The impact on Frame will be lower due to its shared memory. These new patches have already been added to CachyOS, a Linux distribution focused on maximizing gaming performance by integrating cutting-edge features, even if they aren't well tested yet.

          The results can be impressive, with Alan Wake 2 jumping from 14 FPS to 41 FPS on a 4GB GPU, more than trippling the framerate. Other games like RE8 saw more moderate improvements of about 15%, so the impact on something like VRChat remains to be seen.

          One thing the Steam Deck has taught us is that overall the software is more important than the hardware, with Valve coming up with a couple of improvements that allow the Deck to punch way above its weight class. So it's probably not a good idea to extrapolate performance limits from the very resource wasting and not really gaming optimized Windows to SteamOS.

          The hand-ons already showed that Valve apparently cracked the PCVR streaming latency issues with foveated streaming, which uses the same encoders, decoders and Wifi 6e as on Quest 3 PCVR streaming, only in a much more clever way paired with exe tracking.

          Qualcomm described something similar in a white paper several years ago, so anybody could have implemented it. But nobody did until Valve demonstrated it, and by now AVP and Nvidia's CloudXR have added support for foveated streaming. So never underestimate Valve' software engineers. And VRChat no doubt will support ETFR with Frame and Steam machine, reducing render requirements on top of the better VRAM management now available in the Linux kernel.

          On paper, the Steam Machine/Steam Frame combo looks rather weak, but Valve has already proven that they are very good at getting every bit of performance out of weak hardware, and then some. So I'd recomment waiting with statements about a not smooth VRChat experience on the new Valve hardware until after we actually have tests.

          • VR Slut

            Yes, software is more important than the hardware, even more important than subsized hardware with bad software.

  • Christian Schildwaechter

    TL;DR: Valve saying they were trying to keep the price of Frame below that of the Index (full kit at USD 999) was apparently about "a little below", which may now be in question. The post below is mostly speculation about Valve price policy and the resulting much higher expected prices for what was rumored as Deckard.

    While the delay of Steam Frame and Machine seemed inevitable, and those mostly interested in the controller basically begged Valve to release it separately, I'm somewhat concerned about the price. A lot of people were surprised about the USD 99 price tag, and while it certainly is a high quality controller with versatile input options, similarly priced to Sony or Nintendo controllers, and still cheaper than lots of so called Pro controllers, it's still rather expensive.

    The critical part is that it most likely doesn't cost anywhere near USD 99 to produce it. Controllers are high margin peripherals, and when the Oculus Rift CV1 initially shipped with an Xbox One controller that retailed for USD 60, teardowns put its production costs close to USD 15. So about 75% was just margin, something that is very likely also true for the PS Sense controllers Sony offers in Apple stores for USD 299.

    Now Valve said they weren't going to subsidize their new hardware, and justified the price with lots of design and engineering costs that have to be spread over much smaller unit numbers than with Xbox or PlayStation controllers. But it doesn't bode well for their overall pricing policy, as they now sort of set the expectation that what will come next won't be cheap either. They probably could have kept the price of the RAM-less controller lower if they really wanted to. But like with Index, and unlike with the Steam Deck, the name of the game again seems to be high level experience at a high level price.

    Early Deckard rumors mentioned USD 1200, split over two devices, which at the time looked like a bargain for the expected high end specs. We now have learned that there are actually three devices, designed to work together. A "streaming first" HMD offering extremely low latency from a matching mini console offering about six times the HMDs performance, plus a controller that can be 6DoF tracked from the HMD and offers input parity with the Steam Deck for all kinds of flat games to be played on a large virtual screen.

    No doubt the controller would be the the cheapest part, with the HMD most likely the most expensive. After the hand-ons, there was a lot of speculation about esp. the Steam Machine's price, which should perform similar to a PS5 console, so many assumed it would have to be similarly priced to be competitive. Valve engineers reacted with frozen faces when Linus Sebastian told them this in late 2025, hinting that the Machine would actually be more expensive.

    Similarly there were comparisons to the USD 500 Quest 3. Frame will be faster with slightly higher resolution and much better weight distribution, using the very versatile SteamOS with the option to also play flat games, but lacks features like color passthrough, handtracking and Quests large library of mobile optimized games. Frame's performance will be high enough to play Quest games too, but not enough for most PCVR games through local x86 emulation. Leaked framerates in (not yet Frame optimized) HL:A put it at 40-50FPS, so you'd still need a PC for most PCVR games, just like with Quest. Making it hard to justify a significantly higher price.

    Now Valve got lucky/waited long enough so that other vendors raised prices. The cheapest PS5 now costs USD 600, the PS5 Pro USD 900, and the Quest 3 USD 600. But of course Valve's production costs have also risen, so if their devices were planned to be more expensive before, they now will still be.

    So while the rumored pre-RAMpocalypse bundle price might have been for example USD 460 Machine, USD 680 Frame, USD 60 controller to get to USD 1200, by November 2025 this had apparently changed to Machine significantly more than USD 500, Frame close to USD 900-1000, with the controller now about 50% more expensive than many that were hoping for Valve to price things aggressively expected (the 2015 Steam Controller sold for USD 50).

    So what started at USD 1200 for an unspecified Deckard bundle has by now most likely passed USD 1600 and may slowly be creeping towards USD 2000, depending on how long and horrible the AI RAM grab fallout turns out to be. And around the time of the USD 1200 rumor, Deckard was still expected to feature at least 2.5K, if not 4K microOLEDs, and I personally expected it to use an AMD APU to run all PCVR games locally. A 2K LCD HMD paired with a PS5 class mini PC for more than 50% extra on top of that rumored price would be a very different value proposition, no matter how much I'd love to run SteamOS on a standalone HMD.

    • Herbert Werters

      However, the price increases for the PS5 and Quest 3 have not yet reached retailers or customers. Second-hand prices have not risen either. It could take a very long time before these price increases are actually felt by consumers. We do not yet know how they will react to this, however. Therefore, the positive effect for Valve exists only on paper and in theory. People’s purchasing behaviour will show what the real effects will be. Nobody wants dead stock on the shelves.

      Incidentally, it’s already being speculated that Valve may have set the price of the Steam Controller higher in order to subsidise the prices of the Machine and Frame. The Machine will likely come with a controller, and with a price of $99, the Machine then appears cheaper in turn, potentially justifying the higher price to consumers. So, strategically speaking, that could work.

      I feel exactly the same as you about the prices. For me, though, the Frame is still interesting and I’ll be switching from the Quest 3 to the Frame. Just because of the wireless dongle, which you can use really quickly and easily on any computer without cables or a power supply, the controllers with a gamepad layout, and the weight of the HMD and how well-balanced it is on your head. The fact that it uses an open Steam OS with a desktop interface is just brilliant too. Personally, it’s worth it to me. As for everything else, I can also get that after the „RAMpocalypse“.

      • Christian Schildwaechter

        TL;DR: Steam Machine/Frame are strategic products that need large unit sales (for Valve, so more like the Steam Deck than Quest or PS5), and a higher price will inevitably mess this up, with no real option for Valve to salvage this other than accepting (temporary) loss.

        They are screwed, no matter what. The main device would always be the Steam Machine, which was apparently designed to be cheap, reusing unsold stock of AMD mobile GPUs and a custom CPU to lower cost. Valve most likely counted on RAM and flash prices to continue to fall, allowing them to offer it at a rather low price that compensated for the somewhat aged hardware. This would drive sales to casual users and allow to distribute the development costs over more units, again allowing for a lower price.

        This whole plan now went down the toilet. Casual users will more likely wait for one or two years, hoping for prices to drop again. Meaning lower unit sales and less economies of scale benefits, driving up unit costs on top of the already more expensive RAM and flash. Those most likely to be fine with higher prices are Steam fans, but these are also more likely to already have a faster gaming PC.

        The Frame with only 2160p displays was apparently designed to work with the rather pedestrian Steam Machine hardware, as only a small fraction of the existing PCVR library will be updated to use ETFR or other Frame optimizations that might have allowed for a higher resolution.

        Following the Steam Deck playbook, the whole bundle was designed as moderately speced devices that could be offered at an attractive price, which then should lead to lots of unit sales, which would be necessary to establish the Steam Machine as a console-like living room PC. All this falls apart when Valve is now forced to position the same hardware as a high price product, messing up the whole strategy by altering the reachable target group and significantly reducing the unit sales.

        They cannot wait for a year or two either, because even if memory prices should drop back, by then the hardware would be too old to be attractive at a lower price. They have to push the Steam Machine into the market now, so that similar to the Steam Deck the lower performance won't matter that much later due to the by then established great experience, allowing it to remain viable for lets say six years before a successor is released.

        I still want the whole bundle for numerous reasons, but what looked like a very clever plan to get lots of people to switch to console-like PC gaming, and even lure more gamers into VR via flat gaming on virtual screens, now seems to be in danger of failing due to unexpected external factors.

        • Herbert Werters

          Yes, everything you’re saying is spot on. I see it the same way.

          But Valve isn’t the only one facing this problem, and in the current situation, every potential buyer knows exactly why prices are what they are. Unfortunately, all manufacturers have to raise their prices. It’s a bitter pill for everyone to swallow.

          But once the situation returns to normal, prices can be adjusted back to a level where people can afford to buy again. Unfortunately, from that perspective, it just means that all profits are being deferred to the future.

          A war or another crisis that reduces people’s purchasing power would have the same effect. That’s just the way it is, and we have to get through it. We’re not just facing a RAM crisis but also an economic crisis, and people simply aren’t as flush with cash. There are just two issues causing problems at the moment. Everyone has to expect lower sales figures.

          As you say, Valve really ought to rethink its subsidies, at least for the Steam Machine. The Frame is, after all, very niche and certainly not a mass-market product. Yes, the hoped-for hype certainly won’t materialise with such high prices. That’s going to be tricky for Valve. I’m really keen on it too, but I’m also watching my spending.

          • Christian Schildwaechter

            TL;DR: The timing is very important for Valve here, way more than for other companies, and opportunity cost may end up being a bigger issue than memory cost.

            I'm mostly afraid that there is/was a very valuable window of opportunity that could now be missed. Let's be (overly) optimistic and assume that Valve didn't just want to release Steam Machine, Frame and Controller at the same time, but also HL3 as a launch title that would be optimized to run on any combination of these.

            Despite being VR only, HL:A's release made the mainstream press in 2020, drawing a lot of attention to VR. So the reaction to HL3 that has become a meme even non-gamers understand, and that in theory 50 times as many gamers as the 2% VR users on Steam could play, would be even stronger. If Valve launched all of them together, with the Steam Machine offering a not very fast, but affordable low hassle mini PC perfectly suited for the living room, the Controller allowing all games to be played from a couch, and the Frame bringing a both huge and portable virtual screen with high comfort to those without giant TVs, it would kind of create the perfect PR storm.

            Counting on the enormous (flat) Half-Life fan base that by now has been waiting for 19 years for a new release, even lots of non-gaming journalists would pump out articles about them, exploring the prospects of both living room PC gaming and gaming on virtual screens in HMDs in detail, all tested with HL3 guaranteeing high attention and click rates. Valve kind of missed this opportunity in 2019, when HL:A wasn't ready for the Index launch, and instead released almost a year later. Index consequently wasn't covered by the mainstream press, but HL:A still created waves due to the incredible brand value and 13 year delay since HL2 Episode 2.

            And now they may miss it again. If HL3 releases a year after Steam Machine and Frame, it will only be covered as a PC game, not as part of a bundle of new and exciting ways to play it. If they have to launch the hardware at different times, at prices that won't get people to instant-buy them just to get the perfect HL3 experience, and possibly only in limited numbers because Valve cannot get enough memory at bearable prices even when selling at costs, they'll end up losing tons of hardware sales.

            I'm pretty sure HL3 will run on the Steam Deck too, which has been out of stock since February for the same reasons. So launching HL3 now would be a total waste because Valve currently cannot deliver any hardware for playing it. They cannot launch Frame before the Machine, because these are designed as a combo, and a Frame-only release would lead to instant criticism that it cannot run most PCVR games in standalone mode, while this can be explained away by a combined launch with Frame primarily serving as a streaming HMD for Steam Machine. They cannot release the Machine before Frame, because Frame needs the focus on streaming flat games from Steam Machine to draw the attention of the much larger group of non-VR gamers, who'd usually just ignore another VR HMD launching separately.

            The Deck uses LPDDR5, the Frame slightly faster, but otherwise compatible LPDDR5X, so Valve may be forced to choose between using RAM allocations for one or the other. If they do a soft launch, with the Controller first, then Machine and Frame sometimes in later 2026, waiting for the initial sales rush to end to then again start producing more Steam Decks, and only once they are sure they can cover a rush of new sales then finally release HL3, all the momentum and coverage/free PR worth millions that one big, simultaneous launch could create will be diluted away, with only HL3 making mainstream press waves.

            Timing is really important here, and a HL3 plus Steam Machine launch could provide a giant push for all Steam OS hardware, dragging Frame with it further than any VR HMD released by itself could these days. Losing the opportunity for a broad consumer hardware launch alongside an insanely anticipated game may end up a bigger/costlier issue for Valve than the current memory prices. And that's a problem Sony or Meta with established devices, target/user groups and already present in the living room don't have, they can basically sit this out. This situation is particularly bad only for Valve.

          • Herbert Werters

            Yes, that all sounds plausible to me. It really is a dilemma. Ideally, they should wait out the whole RAM crisis, but no one knows how long that will last or whether prices will even fall again. Because once all those AI data centres are built, a great deal of energy will be channelled into lobbying to ensure they are actually used. That’s probably why many people are keen to see gaming and the like move towards streaming. Streaming is going to get a proper boost. I really don’t think this whole development is great.

            By the way, VR is currently taking a real battering. Apple is apparently discontinuing further development of the AVP. This will definitely be seen as a failure in the media and by the public. So VR is really going down the drain, especially as there’s no positive counterpoint at the moment. The launch of the Machine, Frame and HL3 with a VR option would be one, for example.

        • Arno van Wingerde

          And another aspect is that while RAM got massively more expensive, the other components did not… making the difference to e.g. a 4K MicroOLED seem relatively smaller and as the displays get cheaper, the Steam frame will even become less attractive. If prices come down again, a gaming PC can be bought in addition to the headset, for a vastly superior experience for not even that much more money…

          • Christian Schildwaechter

            They are extemely unlucky with their timing. My hope rests a lot on Valve successfully ripping out most of the friction still involved in most VR gaming, and make it an actual pleasure to use, instead of adding a meta/Meta game where you first have to fight with the UI and setup to even get there. Bonus points for a achieving the same with flat games and other apps running in Frame.

            That's pretty much what happened with the Steam Deck, which got dissed a lot before launch for its weak hardware that obviously wouldn't run AAA games properly. This changed rather quickly after the Deck released (and Valve fixed some initial quirks) to more of an "I don't care about the speed, I'll just play games that run well on the Deck, as long as I never have to deal with sh%$#y Windows again." And it actually ran most AAA decently due to a well balanced design with a rather lores resolution that turned out to be just fine.

            Steam Frame and Steam Machine will probaly lose price/performance comparisons against Quest and other PCs, both now and in the future. But (combined) they might win comfort, usability, smoothness of integration and overall experience, which in the end matter more. But like with the Steam Deck, people will first have to have these devices in their own hands before the discussion can switch from paper specs to actual experience.

          • Andrew Jakobs

            But the question of course is, how much displays had they already ordered and how many headsets have already been produced. Because if they haven't produced much and don't have many displays, they might just cancel the current incarnation and wait until prices of ram/storage have come down, which now seems to be a matter of years, not months or quarters, and upgrade all the components when they want to go forward again.

          • Arno van Wingerde

            Sure, but by then we’re talking 2030 or so, Valve does not exactly have a reputation for quick iterations….

      • Andrew Jakobs

        Don't know where you live, but the price increase of the quest 3 is already hitting the stores. Where is was €554 a few weeks ago, it's now €612 in many places..

        • Herbert Werters

          Well, here in Germany, all the Quest models are still available at the old prices. I wonder when that will change. I hope it’s soon, as I’d like to sell my Quest 3 and cash in on the price rise. ;)

    • VR Slut

      The "horrible the AI RAM grab fallout" is lead by one of the one of those moving to Florida to avoid the California billionare tax along with guess who, hyping AI glasses (shmoozing with the top crime families in the world) there!

  • AIvU

    I think I'll hold out for project helix. It's going to be the better device between the two apparently.

  • Christian Schildwaechter

    I think everybody was disappointed by the specs. Those still interested often are because of SteamOS, a lot of them because they own Steam Decks and know how much of a difference it made there.

    We are now close to the fifth anniversary of the Steam Deck announcement, with the hardware performance barely changed since then, and many more powerful handhelds available. Yet even in 2026 the Steam Deck was usually still the most recommend PC gaming handheld (while it was still available). And if you needed something faster, the recommendation for that was a Lenovo Go running Steam OS, or at least installing Bazzite Linux to get close to the Steam Deck's utility and usability.

    So while the new Valve hardware may not be tempting at all for you, given that you already own a lot more powerful devices, a couple of years down the line you might still be temped to run SteamOS on your high end PC and RTX GPU paired with the Pimax Crystal and its eye tracking support to basically emulate a high end Steam Frame/Steam Machine combo. Just like comfort almost always ends up beating cutting edge, the overall experience usually beats specs in the long run.

    • XRC

      Crystal original has it's own XR2 (Octa-core Kryo 585 + Adreno 650 + 8Gb memory) + 256 GB storage onboard would be very interesting to install steam os and run it in standalone mode!

      • Christian Schildwaechter

        I never even considered that option, but it may actually be feasible. After Valve released the FEX x86 emulation for Steam OS ARM, people immediately jumped on that wagon and started to use it/install SteamOS on ARM based gaming handhelds, running x86 Steam binaries on them. Performance is of course limited, but it still works just fine overall.

        It's kind of a wild idea that you may be able to turn your Crystal into an x86 compatible standalone that would allow you to play for example low performance Stardew Valley on a large virtual screen with decent battery life, in addition to serving as a decent streaming target with foveated streaming/rendering. Running any emulated x86 PCVR games on the XR2 Gen 1 might be a challenge though, except for maybe SkarredGhost's "The Unity Cube".

        And given the not exactly huge number of Pimax Crystal sold so far, you'd need some pretty dedicated devs to work on creating a Steam OS distribution that could easily be installed on this HMD. But it would no doubt be a very cool option.

  • NL_VR

    Yes it does, you will see

  • Christian Schildwaechter

    TL;DR: Most Quest games will require basically zero effort to run on Frame; the Frame is way faster mostly due to Valve using a less restricted Snapdragon SoC (plus extra measures to avoid throttling); SteamOS devices are more about the experience than the hardware.

    Meta has shifted Quest to mostly use the cross-platform OpenXR, adding new features there instead of in their own proprietary SDK. In addition most Quest games are created in Unity, with further abstracts platform specific details, so you can basically recompile a Quest game for SteamVR just by picking another target. Quest can run ARM OpenXR APKs that will run the same on a Pico 4, GXR or Frame. If a lot of effort would be required to port a Quest game to another ARM standalone, there'd be basically zero Pico games.

    For Frame, Valve extended SteamOS with Lepton, their version of the Waydroid compatibility layer that, similar to Proton translating Windows calls to native Linux calls, translates Android calls to native Linux calls, allowing for high performance when running foreign OS binaries. So Frame will be compatible with any ARM OpenXR HMD by default. In addition Valve implemented a couple of Meta-specific OpenXR extensions for eye tracking, so your Quest 3 may not support Meta OpenXR apps that use eye tracking, and you may no longer be able to get a Quest Pro that would, but your Frame will use it to gain extra speed in Quest apps from ETFR, making it the fastest Quest you can buy. (All this applies to developers, don't expect users to just be able to copy games directly from Quest 3 to Frame.)

    And you are quite wrong about the performance difference. There may be only one generation of difference between the XR2 Gen 2 and SD8 Gen 3, but XR2 and SD are very different beasts for very good reasons. To start with, the XR2 is produced on an older/cheaper process. This is feasible because its cores are clocked much slower than on the matching SD8 Gen 2, and it only contains performance and economy cores, while the SD8 has additional low power cores and one high performance core.

    This makes sense because phones typically switch between burst and idle use, while VR is a constant load application and never goes into low power mode. So the XR2 using slower cores with moderate power consumption allows it to run at sustained high loads without throttling. But it also means that its peak performance is way lower than that of the matching SD.

    The performance cores in Quest 3 peak at 2.05GHz, but are clocked down to 1.4GHz in graphics heavy apps like games to give the GPU more thermal headroom. This is what allowed for an extra 150% graphics performance boost compared to Quest 2. The performance cores on the SD8 Gen 3 are clocked at 3.15GHz, and it has five of them instead of three on the XR2 Gen 2. The SD8 Gen 3 also uses one high performance core at up to 3.3GHz, though this can only be used for a short time due to its high power draw/heat dispensation.

    All this will give the Frame a way higher peak performance than Quest 3, but also a much higher chance of thermal throttling. Valve's way to deal with this is double sided heat pipes for better cooling, creating very precise models of heat creep inside the SoC, and game specific profiles that adjust the core configurations and clock rates on a per game basis, allowing some to use the high performance for burst calculations, while others may run only on the economy cores to save energy.

    You are also wrong about the XR2 featuring custom blocks. The SD8 already contains the Hexagon DSP that Quest uses for tracking, and the Spectra ISP used for image processing. We don't know if/how Meta uses the Spectra on Quest, or if it is disabled to allow for more power to CPU/GPU/DSP. The XR2 Gen 2 is basically a reconfigured SD8 Gen 2 with some types of cores removed, one extra performance core to compensate for that, a more sustained load friendly configuration, and a bunch of extra ports to connect more cameras, which is the most likely use of the Spectra on Quest.

    We'll have to wait for actual real world tests to really judge the Frame's performance, but just based on what we know about the SoC, your assumptions about lacking performance are most likely very off. It should in fact be significantly faster than even a GXR or Play for Dream, both using an XR2+ Gen 2 with higher clock rates and better cooling than Quest 3. The only standalone faster than Frame will be the AVP.

    And like with the Steam Deck, the big thing about Frame will be Steam OS, not the hardware. Frame will be able to run (some) x86 PCVR games, more PCVR games when they are recompiled for ARM, pretty much every OpenXR game, tons of flat PC and Android games, and every emulated game that runs on the Steam Deck via one of the myriad emulators, with controllers offering full input parity with an Xbox/PS game controller. And that's before even looking at ultra-low latency streaming from a PC. Nobody is bothered with the 1280*800 resolution of the Steam Deck these days, and I'd expect something similar to happen with Frame. This isn't Valve's first VR HMD, portable gaming device, custom controller or Linux based OS with broad compatibility, and they have gotten pretty good at merging all these into a pretty compelling experience.

    • Herbert Werters

      Yes, and that’s because Valve’s business model and core expertise lie in gaming, not social media. It’s worth pointing that out. Meta (aka Facebook) is using it as a platform, and at the moment it looks as though it doesn’t intend to grow any further and Meta is now willing to do less to support it. I won’t go into the possible reasons for that.

      • Christian Schildwaechter

        Given that Valve wins with more people using Steam, I hope that one of the ways they deal with this is (temporarily) stepping back from their statement that the Steam Machine/Frame would be sold with profit, and instead offer it at cost to get things going.

        I don't expect them to subsidize a large hardware roll-out like Meta did, but they can probably affort to compensate some of the current memory price insanity to kickstart Frame/Machine, getting more people onto the living room/virtual console train that longterm could provide them with a lot more money. And hopefully one day memory prices will come down low enough again to get them back to their initial plans.

        • Herbert Werters

          I really hope so. They need to find a way to boost sales and generate more interest. The $99 price tag of the Steam Controller is already taking up a huge amount of space in the coverage. It would be better if the price played a secondary role. If the price overshadows everything else, that’s simply not good. We saw that with the PSVR2. The announcement was really big, but after the price was revealed, it all fizzled out. It’s the same with the Apple Vision Pro. It’s a bit odd that a device like that isn’t selling very well, isn’t it? ;) $3,500 and not much software. Oh well. VR really doesn’t have it easy. So many obstacles in the way. I’ve never seen anything like it.

  • Herbert Werters

    I bet it’ll happen much sooner than that.

    • NL_VR

      Yes you are probably right. It will happen fast

  • Conner

    This is such a stupid move if they wait too long they will lose supporters valve should just get off their asses and release it at what ever price it is at and lower it when the ram shortage gets better

  • Paul Bellino

    The should just simply charge for Half-Life 3 Flat/VR and not bundle it with anything. Charge us 3 times the price for the game and put it towards the price of the Steam Frame and Steam Machine. This way to offset the prices the were raises due to the chip shortage.