Valve today (12 November 2025) announced their new Steam Machine (x86 CPU, 6x more powerful than Steam Deck) and Steam Frame (self-contained and PCVR streaming VR headset with ARM CPU & “FEX” translation of x86 to ARM) to be released in early 2026. No prices yet.
I’m trying to speculate what effects this will have on the wider Linux ecosystem. Both devices will be running Steam OS and be open so you can run any OS.
First, I’ve read many people state that the Steam Deck considerably increased the number of devices running Linux, so it seems to me that these two new devices will accelerate that trend.
Second, it seems to me that the Steam Frame will significantly increase VR use and development for Linux.
Third, I wonder what the implications of Frame’s x86 to arm translation layer (based on FEX, an open source project that I only learned about today) as well as Android compatibility (they state it can sideload Android APKs) will be. Could this somehow help either Linux on Apple silicon or Linux phone efforts? I’m very unfamiliar with what’s going on with either of these efforts, so I may be way out on a limb here.
What do you think about all this?
Edit: this article may prompt some additional thoughts with its discussion of the openness of the Frame - https://www.uploadvr.com/valve-steam-frame-catalog-whole-compatible/
I’ll drop what I said about this in another thread:
I think you probably need to understand the underpinnings of what Valve accomplished over the past few years to understand why the Frame is useful.
Essentially, it’s a Deck strapped to your face. Same OS, same everything, just different hardware platform.
Valve spent the time to revamp SteamOS in order to make it more portable to various devices, which are now launching. Couple that with their efforts on Proton, and you have an entire ecosystem with very little in the way of preventing people from adopting these devices with their ease of use.
Steam Deck was just sort of the appetizer and test launch to gauge interest and build a fully functional hardware development and support vertical in the company, and it was wildly successful. I guarantee (if they can get the price right) that the Frame will sell WAY more units than the awful Vision Pro. I honestly think people might adopt this over buying another version of the Deck if it’s comfortable.
Some things I expect to happen with the Frame launch:
I agree that the opportunity for Frame is to be “big screen” portable gaming.
Desktop stuff will just come along for the ride.
And yes, the ecosystem is in place. Steam is already the de facto distribution channel for games, proton makes most of them work great on Linux, and FEX should make most of those work on Frame.
I am not sure how well FEX works today but it is obviously going to get a lot more love. And the CPU is not the bottleneck for games anyway as the GPU is doing all the heavy lifting.
Sadly not, at least not VR first, cf https://www.uploadvr.com/valve-isnt-currently-working-on-a-new-vr-game/
They also have been tapping Code weavers or just serendipity but there is a lot of movement on the open source x86 to arm translation space too. Which for games on VR headsets is a big deal. https://www.codeweavers.com/blog/mjohnson/2025/11/6/twist-our-arm64-heres-the-latest-crossover-preview
They’ve already said they’ll be using FEX: https://www.pcgamer.com/hardware/vr-hardware/steamos-launching-for-arm-fex-translation-layer/
Interesting comments, thanks!
I fully agree that this will sell way more than Vision Pro. I think this is pretty much guaranteed. The highest price I’ve seen estimated for the Frame is $1200, so it will much cheaper and much more versatile.
I also think that Theater mode for media is pretty much a guarantee at release, given that they’ve already demoed playing regular non-VR games in Theater mode.
I’ve also seen some mentions of Linux desktop on it, but haven’t seen any concrete details about it.
To be fair, I think they’re slightly different markets.
The AVP is a “Productivity” device and seems more focused on “Mixed Reality” use with it’s super high quality passthrough and what not, vs the Frame which is more focused on gaming, has black and white passthrough and (I’m assuming) no Real World mapping so you can’t have floating windows that stay where you put them in your physical space for example.
That being said I think more sales than the AVP is a guarantee, on the price alone.
If anything, the real question is if that “VR Productivity” market that Apple is targeting really exists. (didn’t the HoloLens fail?)
That’s true about them being different markets and also the comparison to HoloLens!
Where have you heard that Frame won’t have Real World mapping?
That was just an assumption, I should probably make that clearer on my comment.
Got it, thanks for the clarification. I would be surprised if any VR headset with inside-out tracking wouldn’t have real world mapping today, but we’ll see!
Desktop is a built-in feature of SteamOS. They’d have to actually work to remove that by default. No reason for it not to be there.
Yeah, but I don’t think KDE has VR capabilities. So it’ll be interesting to see how that’ll work. They mentioned the ability of opening desktop applications in VR. So I think you’ll be able to position those in space.
It wouldn’t need to. They would just need to include a virtual desktop manager or interface to render the usual compositor in a VR sort of way. That’s why I put it in the list. Same thing that would make a theater mode would also allow a desktop to render in a space on a VR compositor.
Yeah, I can understand that. I had heard that Steam Deck had a desktop feature built-in. Some of the videos and articles about the Machine have shown and mentioned desktop apps and KDE, but not regarding the Frame, so I wasn’t sure, especially considering that Frame will be using a different hardware architecture.
That’s already a thing, I saw a video of someone streaming a VR game to a Quest 3 via Steam Link and using it to map hand gestures to regular VR controller actions, allowing him to play games with no hand tracking support controller-free
That’s slightly different. That is mapping controllers to already existing inputs for a game, which steam-input already does.
Mapping all the sensors of a VR headset for motion and tracking is an entirely different thing, though kinda similar in some sense.
But, what’s the user experience playing the MYST humble bundle on it?
Well it’s not going to suddenly be all VR’d up or anything 🤣
Part of the reason I would imaging they implementing a new kind of steam-input layer for VR is for things like a theater mode and desktop. I could see them making some sort of a simple hook for view controls in games for your exact scenario, but that would be heavily dependent on the game having something like free look already be possible, and then the developers just write a quick patch of a couple lines to hook the steam-vr-input hook into their code, and BAM.
I’m just wondering about basic functionality - can it play through without a crash or is it still Windows only for a smooth UX?
99% of all Windows games running Proton, and most perform better than on Windows, depending othe game.
For the specific games you mentioned, they all have Platinum rating in Proton, meaning flawless. You can see here in ProtonDB.
I’m not sure what your experience was in the past, but I write tons of Proton patches for games, and the only ones I’ve seen that don’t play well are the pre-DirectX9 games, which can’t plan on Windows XP or later anyway. Proton will soon be able to play these games without issue thanks to some Vulkan patches coming up.
…. and all your credibility is gone.
You don’t read much, eh? Valve has confirmed it, people have seen it, and multiple artists and devs have confirmed working on the finishing touches in the past year.
They released Alyx with the Valve Index, which also confirmed to be a drop-in VR engine for the Half-Life games. Valve likes releasing their IP games as surprises, and no better way to sell out millions of units than releasing HL3 along with this new hardware.
Seems like a pretty safe bet.
That makes a lot of sense to me. I didn’t realize that there was already evidence for a new Half-Life existing out there.