When Valve shipped the Steam Deck in February 2022, it did two things simultaneously: it put a polished handheld gaming PC in consumers' hands, and it confirmed that a fully realized, Arch-based, gaming-first Linux operating system -- SteamOS 3 -- existed and worked. The problem was that Valve never released it as a general-purpose installer for desktop hardware. The project codenamed Holo stayed locked to the Deck's recovery image, available to everyone to inspect but not to cleanly install on their own machines.

HoloISO was the community's answer to that gap. Built by a developer known as theVakhovskeIsTaken, it repackaged the Steam Deck's SteamOS components into a bootable installer that could run on standard x86_64 hardware. For two-plus years, it was the closest thing available to "real" SteamOS on a gaming desktop. Then it ended -- and the story of why it ended tells you almost as much about the state of Linux gaming as the project itself did.

This is a story about the gap between when a platform becomes technically possible and when it becomes institutionally available — and what lives in that gap. When Valve built SteamOS 3 for the Steam Deck, they accidentally created the compelling Linux gaming platform that had been theorized for years. They just did not ship it to everyone. HoloISO existed in that gap: a community-built proof of concept that the future was already here, just unevenly distributed.

Understanding HoloISO means holding three things in mind at once: the technical architecture that made it work, the organizational constraints that ended it, and the way it validated demand that larger and better-resourced projects then stepped in to serve. That pattern — community innovation preceding platform expansion — is the central story of Linux gaming in the 2020s, and it keeps repeating.

Why SteamOS 3 Was Worth Chasing

To understand what HoloISO was trying to deliver, you need to understand what made SteamOS 3 different from anything that had come before it. Valve had previously shipped SteamOS 2, a Debian-based operating system built for the Steam Machine initiative that launched in 2015 and collapsed shortly after. That version still technically exists on Valve's servers with a warning that it is not compatible with the Steam Deck.

SteamOS 3 is architecturally different. As Tom's Hardware noted when HoloISO first appeared, the new version abandoned Debian entirely in favor of Arch Linux as its foundation, and introduced Proton for Windows game compatibility rather than relying on a small native Linux library. It also introduced a custom micro-compositor called $ man gamescopeValve's micro-compositor that runs each game in an isolated Xwayland sandbox. Handles resolution spoofing, FSR upscaling, frame rate limiting, and direct DRM/KMS frame delivery — bypassing the desktop compositor for minimum latency. to serve as the session layer managing how games are presented on screen.

Tom's Hardware, writing in May 2022, characterized Valve's decision not to release a public SteamOS 3 installer as a notable oversight -- made more conspicuous by the fact that the outdated Debian 8-based SteamOS 2 remained available on Valve's own servers, complete with a compatibility warning.

The combination was significant. Arch Linux's rolling release model meant the system could stay current with kernel and Mesa updates without requiring full OS upgrades. Developed in collaboration with CodeWeavers, the $ man protonValve's compatibility layer built on Wine, DXVK, and VKD3D-Proton. Translates Windows game API calls (Direct3D, Win32) to Vulkan and Linux equivalents. Enables approximately 90% of Windows Steam games to run on Linux without modification. compatibility layer translated Direct3D calls to Vulkan in real time using the $ man dxvkA Vulkan-based translation layer for Direct3D 9, 10, and 11. One of the key components inside Proton. Responsible for the dramatic performance improvements in DX11 games on Linux, often matching or exceeding Windows equivalents on AMD hardware. translator for Direct3D 8 through 11, and the $ man vkd3d-protonValve's heavily patched fork of VKD3D — translates Direct3D 12 to Vulkan. VKD3D-Proton 3.0 (November 2025) added AMD FSR 4 and Anti-Lag 2 support, extending those to older RDNA GPUs. translator for Direct3D 12. The result was that the vast majority of Windows games in a Steam library could simply run -- no configuration, no manual Wine prefixes, no per-game setup rituals.

Gamescope added another layer of capability. According to the project's own documentation, it runs each game inside its own isolated Xwayland sandbox, handles resolution spoofing so a game sees exactly the screen geometry you want it to see, upscales output using AMD FidelityFX Super Resolution, and can flip frames directly to the display via DRM/KMS when compositing is not needed -- eliminating latency that a traditional compositor would introduce. For gaming, this architecture is meaningfully better than running Steam inside a standard desktop environment.

What HoloISO Actually Was

HoloISO's stated goal was to bring the Steam Deck's SteamOS Holo redistribution into a generic, installable format and provide a close-to-official SteamOS experience. It was not a fork in any deep sense -- the developer described it as attempting to be 99% of the way to SteamOS, with the code and packages sourced straight from Valve and the ISO built using the same rootfs bootstrap as all HoloISO installations.

The primary engineering challenge was not replicating what Valve open-sourced. It was reimplementing the proprietary components that assumed they were running on Steam Deck hardware. The Steam client, Gamescope, and various user-created Deck applications expected certain hardware abstractions that do not exist on a generic desktop. HoloISO had to stub or reimplement those components -- things like the BIOS update helper, the SteamOS session selector, and the hardware-specific polkit rules -- so that the software stack believed it was operating in a valid environment.

Architecture Note

HoloISO shipped Valve's own pacman.conf repositories alongside its own holoinstall script and post-installation binaries. On first boot, users were greeted with the Steam Deck's out-of-box experience screen, from which they could log into Steam and immediately enter Gaming Mode, or switch to KDE Plasma desktop mode via the power menu -- identical to the Deck experience.

Installation was straightforward by Linux standards. Users wrote the ISO to a USB drive using BalenaEtcher, Rufus in DD mode, or a manual dd command, booted from it, and followed the installer. The resulting system booted directly into Gamescope and the Steam Gaming Mode interface.

flash HoloISO to USB
$ sudo dd if=SteamOS.iso of=/dev/sdX bs=4M status=progress oflag=sync
# Replace /dev/sdX with your actual USB device -- use lsblk to confirm
# BalenaEtcher or Rufus (DD mode) are safer options if unsure

What It Could and Could Not Run

In practice, HoloISO worked best on AMD hardware and was essentially unsupported on NVIDIA. This was not an arbitrary restriction -- Gamescope itself is designed to run on Mesa with AMD with minimal effort, and the proprietary NVIDIA driver stack has historically had friction with Valve's Wayland-based compositor. The project documentation was blunt about this: NVIDIA GPU users were told they were on their own, as Valve's own Steam client updates had broken the Game Mode UI on NVIDIA systems and no support would be provided.

Intel integrated graphics was a middle ground -- possible, but requiring Mesa and Gamescope configuration work, and described as hit-or-miss in practice. For AMD users, both APUs and dedicated cards, the experience was generally solid. Games in a Steam library that carried Deck Verified or Deck Playable badges worked as expected. Proton handled titles ranging from the simply old to the recently released.

Community sentiment on Steam discussion boards at the time reflected this consistently: users running all-AMD hardware generally reported a solid, console-like experience and recommended the project for that use case specifically.

Real-world testing documented on independent blogs confirmed this assessment. A system built around an AMD Ryzen 7 5700G APU -- which provides roughly comparable GPU performance to the Steam Deck at higher power limits -- ran games like BeamNG.Drive and Dirt Rally 2 smoothly. GTA V, including GTA Online, installed and launched without issues. Older titles presented more variation, and games with kernel-level anti-cheat remained the persistent problem they are on all Linux gaming platforms.

The anti-cheat wall is worth addressing directly because it is the most common objection raised against Linux gaming in general. Games using BattlEye and Easy Anti-Cheat can work on Proton when publishers explicitly enable support for their builds -- this requires a manual step on the developer or publisher side, but is technically feasible and an increasing number of games have completed it. Kernel-space anti-cheat solutions, however, fundamentally cannot run inside the Wine environment and remain incompatible. Titles like Fortnite, Apex Legends at various points, and several battle royale games with aggressive kernel anti-cheat have been blocked specifically because of this architecture, not because of Proton's broader capability.

The End of the Original HoloISO

In January 2024, the original HoloISO GitHub repository was archived. The developer posted a public service announcement declaring the original project end-of-life and directing users to a new, immutable version of HoloISO. The archived repository is now read-only -- no updates, no issue support.

GamingOnLinux reported that adopting an $ man immutable-osAn OS design where the root filesystem is read-only. Updates replace the entire root image atomically rather than patching individual packages. Prevents most classes of update-related system breakage. Used by SteamOS, Fedora Silverblue, and the immutable HoloISO. Trade-off: users cannot freely install system packages with traditional package managers. design was the primary motivation. The original HoloISO was a mutable Arch-based system -- users could install packages with pacman, modify system files, and break things the way you can break any traditional Linux install. The immutable version adopts the model used by SteamOS itself: a read-only filesystem that gets atomically replaced on updates, improving stability, security, and update reliability. The developer stated that the immutable version was ready for daily use.

As GamingOnLinux summarized the developer's explanation: the new design follows the model used by SteamOS and Fedora Silverblue, keeping the core system read-only and replacing it atomically on updates -- a change intended to improve security, stability, and update reliability compared to the original mutable Arch install.

Case Study: When Community Infrastructure Becomes a Dependency

The AYANEO Incident and What It Exposed

The timing of HoloISO's original EOL had an immediate real-world casualty. Hardware manufacturer AYANEO had planned to ship the AYANEO NEXT LITE with HoloISO as its operating system — a commercial device built on a one-person community project. When the EOL announcement arrived, abruptly, without a public migration path for existing installations, AYANEO was caught mid-production with a product whose OS foundation was being discontinued.

The company pivoted back to Windows rather than attempt a rapid transition to the immutable version. The NEXT LITE shipped, but not as the Linux-native device it was designed to be.

This is the risk the transparency concerns (detailed in the callout immediately below) were pointing toward. A project distributed through a personal Telegram channel, without reproducible builds, without an organizational structure, can disappear as quickly as it appeared — and any commercial dependency built on it disappears with it. HoloISO was never designed to be infrastructure. AYANEO treated it as if it were. The lesson was not that community projects are unreliable; it was that the criteria for evaluating reliability are different when a commercial product is depending on someone else's open-source side project.

Project Transparency Concerns

Beyond the EOL event, independent observers noted structural problems with how HoloISO was run. A detailed 2025 analysis by developer Bret.io noted that HoloISO was distributed primarily through a Russian Telegram channel and had build and distribution practices that were difficult for outside contributors to verify. The project was not structured to allow reproducible builds, making it hard to confirm that the distributed ISOs matched the source code. While there is no documented evidence of malicious code, the opacity was a legitimate concern for any production deployment.

The Immutable Successor and What It Means

The immutable HoloISO operates on the same principle as SteamOS itself. The system root is read-only. Updates replace the root image atomically rather than patching individual files. This prevents the most common class of system breakage on traditional Linux -- a half-applied update, a conflicting package, a misconfigured dependency. When an update fails partway through, the old image is still intact and the system can roll back.

For a gaming-only machine, immutability is largely a net positive. The trade-off is that users cannot freely install system-level packages with pacman. Applications are expected to arrive as Flatpaks or through other containerized formats. This constraint is the same one SteamOS itself imposes, and most gaming use cases operate entirely within the Steam and Flatpak ecosystems anyway. Power users who want a system-level Arch install with Gamescope bolted on will find the immutable model restrictive; users who want a gaming appliance will not notice.

Is HoloISO Still Worth Installing in 2026?

The honest answer is: probably not, unless you have a specific reason to use it. The community has largely moved toward Bazzite, which has a larger development team, cleaner build infrastructure, broader hardware support including NVIDIA, and active maintenance. HoloISO's immutable version remains available via github.com/HoloISO, but new users starting fresh in 2026 will find Bazzite a smoother experience.

Desktop Mode and Day-to-Day Use

Gaming Mode gets most of the attention in coverage of SteamOS-derived systems, but a meaningful portion of users also need a functioning desktop environment -- for a browser, a chat client, a file manager, or any task that a 10-foot couch interface handles poorly. HoloISO, Bazzite, and SteamOS itself all ship KDE Plasma as the desktop environment accessible through the power menu's "Switch to Desktop" option. The experience is largely what you would expect from KDE Plasma on any Arch or Fedora system: functional, fast, reasonably polished, and occasionally in need of manual configuration to work well with gaming-specific hardware.

On SteamOS and the original HoloISO, the desktop is where the immutable model starts to show its edges for non-gaming tasks. System packages are read-only, which means you cannot install applications via the system package manager the way you would on a conventional distribution. The intended workflow is Flatpaks from the Discover software center, which covers a wide range of common desktop applications -- Firefox, VLC, LibreOffice, Discord, Spotify, and most other common tools are available. For software not packaged as a Flatpak, the system provides a distrobox utility that creates containerized Linux environments with full package manager access. It works, but it is a layer of abstraction that will feel unfamiliar to users expecting a traditional Linux desktop.

Bazzite inherits and extends this model. The ujust command-line utility provides convenient shortcuts for common setup tasks -- installing NVIDIA drivers, configuring peripherals, enabling Waydroid for Android app support -- that would otherwise require manual research on a new installation. For most desktop gaming needs, the Flatpak ecosystem plus distrobox covers the gap. For power users who want to run background processes, custom scripts, or system-level software that is not containerizable, the constraint is real and worth factoring into the decision.

What Desktop Mode Is and Is Not

Desktop mode on SteamOS and its derivatives is not a full traditional Linux desktop. It is a KDE Plasma session that shares the same read-only root with the gaming environment. You can browse the web, run Flatpak apps, and use Steam as a standard desktop application. You cannot install system packages with pacman or dnf without workarounds, and changes outside your home directory do not persist across system updates. For gaming appliance use cases, this is a non-issue. For users who need a daily-driver workstation, it is a real constraint -- and a sign that a general-purpose distribution with Gamescope bolted on may serve better than a gaming-first immutable OS.

The Broader Linux Gaming Ecosystem in 2026

HoloISO's trajectory from community innovation to EOL to replacement is not an isolated story -- it reflects the broader maturation of the Linux gaming ecosystem. When HoloISO launched in 2022, the options for a SteamOS-like desktop experience were thin. By early 2026, the landscape has changed substantially.

May 2022
Community
HoloISO launches
theVakhovskeIsTaken releases first bootable SteamOS 3 installer for generic x86_64 hardware. Valve has no equivalent public release.
2022–23
Community
HoloISO matures; Bazzite and ChimeraOS emerge
The SteamOS-like distribution space grows competitive. Bazzite distinguishes itself with NVIDIA support and Fedora Atomic base. ChimeraOS develops its own independent lineage.
Jan 2024
EOL
Original HoloISO archived; immutable successor announced
Mutable Arch-based HoloISO becomes read-only. AYANEO pivots the NEXT LITE back to Windows. An immutable replacement is announced but the ecosystem has started moving toward Bazzite.
May 2025
Valve
SteamOS 3.7.8 — first official third-party hardware support
Valve officially supports Legion Go S (pre-installed), Legion Go, and ASUS ROG Ally. The demand HoloISO validated in 2022 is now an official Valve product line.
Nov 2025
Hardware
Steam Machine announced; Proton 10.0-3 released
Valve reveals Steam Machine, Steam Controller redesign, and Steam Frame VR headset. Linux reaches 3.58% of Steam in December — an all-time high. Proton 10 ships with dozens of game-specific improvements.
2026
Valve
Steam Machine delayed; still confirmed for H1 2026
RAM and component shortages delay launch. At GDC March 2026, Valve confirms active RAM sourcing. "Powered by SteamOS" partner device program expands. Anti-cheat commercial pressure now explicitly part of Valve's strategy.

Bazzite

Bazzite is currently the community consensus recommendation for a gaming-focused Linux distribution on non-Deck hardware. Built on Fedora Atomic Desktops rather than Arch, it uses the same Gamescope compositor and the same Steam Gaming Mode interface as SteamOS, but with a wider hardware support surface. NVIDIA GPU support works out of the box with proprietary drivers baked in. AMD and Intel are also supported. The project's own documentation describes it as including features SteamOS lacks, including the System76 CPU scheduler, Waydroid for Android app support, and HHD for fine-grained power management on handheld hardware.

The critical distinction is that Bazzite is not SteamOS. It is a Fedora-based distribution that borrows heavily from SteamOS's components and aesthetic but has its own lineage. Under the hood, package management, update mechanisms, and the base system behave like Fedora, not Arch. This matters when troubleshooting or tinkering.

ChimeraOS

ChimeraOS predates modern SteamOS and has its own history. It boots into Gamescope and Steam Gaming Mode as its primary interface and is immutable. Unlike HoloISO, it is based on GNOME for its desktop mode rather than KDE Plasma. The ChimeraOS creator has described the project as hyper-focused on the console gaming use case, with the desktop environment added only because users wanted it -- not as a primary design goal.

Non-Steam Game Libraries: GOG, Epic, and Heroic

The article has so far focused on Steam because Steam is the infrastructure on which all of these distributions are built. But a significant portion of PC game libraries live outside Steam -- on GOG, Epic Games Store, and other storefronts. HoloISO, Bazzite, and SteamOS address this through Heroic Games Launcher, a Flatpak-packaged open-source client that supports both GOG and Epic libraries with Proton compatibility built in. Lutris serves as a more general-purpose launcher for older games and emulators. Neither is as seamless as the Steam integration -- there is no automatic Proton version selection, no equivalent of the Deck Verified status, and Gamescope integration requires manual launch options -- but the capability exists and works for a substantial portion of the non-Steam catalog. GOG's DRM-free titles tend to fare particularly well, since they carry no additional compatibility layers beyond Proton itself. Epic-exclusive titles with kernel anti-cheat, such as Fortnite, face the same wall discussed in the anti-cheat section regardless of which launcher delivers them.

SteamOS Itself

Valve has made significant moves to expand SteamOS beyond the Steam Deck. In May 2025, SteamOS 3.7.8 -- internally codenamed "Go Country" by Valve -- became the first stable build to officially support third-party hardware: full support for the Lenovo Legion Go S (the first partner handheld to ship with SteamOS pre-installed), and expanded support for the Lenovo Legion Go and ASUS ROG Ally. NotebookCheck confirmed the 3.7.8 release requires AMD CPU and GPU hardware with an NVMe SSD, which means Intel-based handhelds such as the MSI Claw remain unsupported. The update also brought Linux kernel 6.11, KDE Plasma 6.2.5, and updated Mesa graphics drivers. Valve stated it is working with additional partners on officially licensed "Powered by SteamOS" devices. The AMD requirement also explains why NVIDIA desktop hardware is not yet viable.

The broader desktop release -- for generic PC hardware -- remains incomplete. Valve developer Pierre-Loup Griffais explained at CES 2025 that Intel and NVIDIA driver maturity are the blocking factors. GamingOnLinux reported his statement directly: driver support on some platforms remains at a basic level, with Intel work ongoing and NVIDIA's open-source driver integration still requiring significant development before either can anchor a broad SteamOS release. Valve is not targeting Windows market share -- Griffais was explicit that it is not a goal to convert users who already have a good experience on Windows -- but is building SteamOS as a meaningful alternative for users who want it.

Linux Gaming by the Numbers (as of December 2025)

Linux reached 3.20% of Steam's active user base in November 2025. Valve later revised the December 2025 figure upward to 3.58% — an all-time high, representing approximately 4.7 million monthly active gamers. SteamOS Holo powers 26.4% of Linux gaming installations, driven by Steam Deck sales. Proton enables approximately 90% of Windows Steam games to run on Linux. Valve's Deck Verified program has certified over 21,694 titles as playable on SteamOS as of November 2025. AMD CPUs power roughly 70% of Linux gaming systems — a figure driven heavily by the Steam Deck's custom AMD APU. Source: commandlinux.com Steam survey analysis and Phoronix.

Proton and Gamescope: The Real Foundation

Any discussion of Linux gaming -- whether on HoloISO, Bazzite, or SteamOS itself -- ultimately comes back to Proton and Gamescope, because these are what actually run the games. The distribution is largely a delivery mechanism for these two components.

Proton's development has continued at a rapid pace. Proton 10.0-3, released November 13, 2025, brought a broad set of per-title improvements across a long list of specific games, including fixing crackling audio in Metal Gear Solid V: The Phantom Pain cutscenes on Steam Deck, resolving a login page issue in Crysis 3, improving performance in Rocket League, and addressing five-minute freezes in Deadlock. The release was developed in collaboration with CodeWeavers and represents the current stable branch of the compatibility layer that enables tens of thousands of Windows games to run on Linux. One week later, VKD3D-Proton 3.0 was released with AMD FSR 4 and Anti-Lag 2 support, extending those capabilities to older RDNA GPUs beyond the hardware that ships with native FSR 4 support.

Proton is not a single tool — it is a stack of components, each addressing a different compatibility layer between Windows game APIs and Linux graphics infrastructure. Here is what each layer does:

Proton compatibility stack
Game
Windows x86_64 executable — calls DirectX or OpenGL APIs, unaware it is running on Linux
DXVK
Translates Direct3D 9, 10, and 11 calls to Vulkan. Responsible for the performance gains that closed the gap with Windows on DX11 titles.
VKD3D-Proton
Translates Direct3D 12 calls to Vulkan. Valve's heavily patched fork of the upstream VKD3D project. VKD3D-Proton 3.0 added AMD FSR 4 and Anti-Lag 2 in November 2025.
Wine
Runs the Windows executable itself — translates Win32 system calls, file system paths, and process management to Linux equivalents. Proton's Wine carries patches (esync, fsync) that reduce CPU synchronization overhead not in upstream Wine.
Vulkan / Mesa
The native Linux graphics API and open-source GPU drivers. AMD hardware (RADV driver in Mesa) provides the most mature path; Intel Arc follows; NVIDIA's open-source NVK is still maturing.

Proton's architecture layers several components. DXVK translates Direct3D 8, 9, 10, and 11 calls to Vulkan. VKD3D-Proton handles Direct3D 12. Both sit on top of a heavily patched version of Wine. Additional patches for synchronization primitives -- esync and fsync -- reduce CPU overhead in games that make many synchronization calls. The result is that as Wikipedia's Proton article notes, Proton generally has better compatibility than upstream Wine due to these additional patches and components that Wine itself has not accepted into its main codebase.

Gamescope's role is equally important but less visible to the end user. Running inside Gamescope, a game cannot interact with or interfere with the desktop environment, cannot crash the compositor, and cannot cause resolution or refresh rate issues that persist after it closes. The game believes it is the only thing on screen because, inside its Xwayland sandbox, it effectively is. Frame delivery uses async Vulkan compute, which means the display receives a new frame quickly even when the GPU is busy rendering the next one -- a latency reduction that matters especially in competitive or timing-sensitive games.

gamescope launch examples
# Upscale a 720p game to 1440p using integer scaling
$ gamescope -h 720 -H 1440 -S integer -- %command%

# Cap a vsync'd game to 30 FPS
$ gamescope -r 30 -- %command%

# Run at 1080p but output to 3440x1440 ultrawide (pillarboxed)
$ gamescope -w 1920 -h 1080 -W 3440 -H 1440 -b -- %command%

# Use AMD FSR upscaling
$ gamescope -F fsr -h 720 -H 1080 -- %command%

Performance Compared to Windows

One of the questions that prospective Linux gaming converts ask most frequently — and that most SteamOS coverage leaves unanswered — is whether gaming performance on Linux is worse than Windows. The answer in 2026 depends heavily on your hardware and the specific game category. Rather than a blanket verdict, here is the actual decision logic:

Performance decision framework — Linux vs. Windows (2026)
Single-player, AMD GPU, Steam library — Proton with Vulkan backend (DXVK/VKD3D-Proton)
Negligible gap. Some titles match or exceed Windows. Not a reason to choose Windows.
First launch of a new game — shader cache not yet compiled
Expect stutter for 5–20 minutes. Diminishes on repeat play once cache warms. Not a performance deficiency — a one-time cost.
Multiplayer with BattlEye or EAC — publisher has enabled Proton support
Works. Performance comparable to AMD Windows gaming. Check ProtonDB for per-title confirmation.
Multiplayer with kernel-mode anti-cheat — Fortnite, Apex (certain builds), Valorant
Blocked. Kernel anti-cheat cannot run inside the Wine environment. This is architectural, not fixable by Proton updates.
NVIDIA desktop hardware — any SteamOS-derivative distro
Use Bazzite specifically — it ships proprietary NVIDIA drivers. SteamOS and HoloISO do not support NVIDIA reliably. Expect occasional Gamescope friction.
Intel Arc GPU
Functional but behind AMD in driver maturity and Gamescope stability. Intel-based handhelds (MSI Claw) lack official SteamOS support entirely.
OpenGL titles, Denuvo DRM
OpenGL translation overhead is higher than Vulkan. Denuvo adds overhead on top of Proton's translation layer. Noticeable but not a dealbreaker for single-player play.

For games that run through Proton with Vulkan backends (DXVK for DX9-11, VKD3D-Proton for DX12), the performance gap relative to native Windows has narrowed to the point of being negligible in many titles, and in some cases Linux performance exceeds Windows. This is not a theoretical observation -- community benchmarking across titles like Cyberpunk 2077, Elden Ring, and Baldur's Gate 3 consistently shows AMD hardware on Proton matching or approaching Windows frame rates, particularly after shader cache warm-up removes the first-launch stuttering that gives Proton an unfair reputation in early benchmark runs.

The situations where Linux meaningfully underperforms Windows fall into recognizable categories. First-time shader compilation produces stutter that does not appear on Windows, because DirectX 12 games manage their own shader caches and Windows already has months of pre-compiled shaders in circulation before a Linux user first launches the title. Proton's shader pre-compilation partially addresses this, but does not eliminate it. OpenGL games have historically had more overhead under translation than Direct3D equivalents, though this is a shrinking category. Some titles with aggressive DRM -- Denuvo, in particular -- add overhead on top of Proton's existing translation layer, producing worse relative performance than the same titles without it.

On Intel hardware, the performance picture is less favorable. The Intel Arc GPU ecosystem on Linux has matured but remains behind AMD in driver stability and Gamescope integration -- a factor that shows up in benchmarks and in the absence of official SteamOS support for Intel-based handhelds like the MSI Claw. AMD hardware remains the reference case for Linux gaming performance expectations.

Practical Benchmark Reality

For a single-player game library on AMD hardware, Linux gaming performance in 2026 is close enough to Windows that the difference is rarely the reason to choose one platform over the other. The relevant questions are usually anti-cheat compatibility, specific title quirks, and whether the immutable OS model suits how you use your machine. Performance is not the barrier it was in 2018 or 2020.

The Persistent NVIDIA Problem

One thread that runs through every SteamOS-derivative project is the complicated relationship with NVIDIA hardware. SteamOS itself, HoloISO, and the original gamescope-session setup all work best -- or exclusively -- with AMD. Bazzite is the significant exception, having invested engineering effort in making NVIDIA work.

The root cause is architectural. Gamescope uses DRM/KMS to flip frames directly to the display, bypassing the compositor stack when possible. NVIDIA's proprietary driver has historically had limited support for the GBM (Generic Buffer Management) backend this approach requires on Linux. Valve developer Pierre-Loup Griffais confirmed this directly in a January 2025 interview reported by GamingOnLinux: driver support on both Intel and NVIDIA platforms remains at an early stage, with NVIDIA's open-source driver integration still requiring considerable work before it could underpin a reliable SteamOS release for general PC hardware.

The situation is improving incrementally. The Mesa graphics stack now includes an open-source $ man nvkThe open-source NVIDIA Vulkan driver being developed inside the Mesa graphics stack. Represents the long-term route to broad NVIDIA Linux compatibility without the proprietary driver. Still maturing as of early 2026. Vulkan driver for NVIDIA hardware, representing the long-term route to broad NVIDIA compatibility without dependency on the proprietary driver. HoloISO's immutable version added NVK testing documentation, with a disclaimer that no official support is provided. For users with current NVIDIA hardware who want a SteamOS-like experience today, Bazzite with its bundled proprietary driver support is the only practical choice.

Hardware, Handhelds, and What Comes Next

The Linux gaming ecosystem in 2026 is no longer primarily a desktop story. The Steam Deck normalized Linux gaming for consumers who would never have chosen a Linux distribution voluntarily. According to commandlinux.com's Steam survey analysis, the Steam Deck had sold over 5.6 million units by mid-2025, with some community estimates placing the figure closer to 8 million by year end. Every unit runs SteamOS, making it the single largest vector for new Linux gaming installations by a wide margin -- SteamOS Holo alone accounts for over 26% of all Linux installations on Steam.

The category has expanded decisively beyond Valve. In May 2025, SteamOS 3.7.8 officially extended support to the Lenovo Legion Go S, Lenovo Legion Go, and ASUS ROG Ally -- the first time Valve provided sanctioned installation instructions for non-Deck hardware. NotebookCheck confirmed the Legion Go S launched with SteamOS pre-installed, priced significantly lower than the equivalent Windows model. Valve stated it is working with additional partners on officially licensed "Powered by SteamOS" devices. The AMD-only requirement currently excludes Intel-based handhelds such as the MSI Claw, and NVIDIA desktop hardware remains unsupported.

Hardware-specific compatibility work -- TDP control, button mapping, display rotation, gamepad virtual controller support -- has become a focus of Bazzite and SteamOS itself in ways HoloISO never prioritized, because HoloISO was built for desktop hardware with conventional input. The handheld category in 2026 is competitive: ASUS, Lenovo, Ayaneo, GPD, and others all ship capable hardware, and the question of operating system has become a genuine differentiator rather than an afterthought.

Engadget's ongoing coverage of Valve's 2026 hardware lineup documents that Valve announced the new Steam Machine on November 12, 2025, alongside a redesigned Steam Controller and the Steam Frame wireless VR headset. The Steam Machine is a compact small-form-factor desktop running SteamOS, powered by a semi-custom AMD Zen 4 six-core CPU (up to 4.8GHz, 30W TDP) and a discrete AMD RDNA 3 GPU with 28 compute units and 8GB GDDR6 VRAM -- a configuration close to the mobile RX 7600M, drawing 110W and delivering, according to Valve, roughly six times the performance of the original Steam Deck. RAM is 16GB DDR5 via replaceable SO-DIMM modules, with 512GB or 2TB NVMe storage options. In February 2026, Valve confirmed a delay to the release timeline, citing RAM and component shortages, while affirming its goal of shipping in the first half of 2026. At GDC in March 2026, Valve stated it was still actively sourcing RAM. Game Rant reported Valve subsequently updated its official page to confirm it will "ship all three products this year." No price has been announced. At GDC 2026, Valve also shared that Steam Machine Verified games will be expected to support the same input methods as the Steam Deck and run at 1080p at 30fps at minimum -- a requirement that Deck Verified titles automatically satisfy. These will be the first Valve-branded non-handheld devices to ship with SteamOS 3, and their existence is expected to accelerate Valve's investment in broader hardware compatibility -- while also putting meaningful commercial pressure on publishers to enable anti-cheat support.

Anti-Cheat: The Remaining Wall

The single most consistent limitation of Linux gaming -- across HoloISO, Bazzite, native SteamOS, and any other platform -- is kernel-mode anti-cheat. A February 2026 analysis on WindowsForum argued that publisher decisions, not technical barriers alone, are what keep certain games unsupported on Linux. BattlEye and Easy Anti-Cheat both have technical paths to Proton compatibility when publishers enable support. The barrier is not engineering -- it is publisher willingness to complete the enabling steps.

Valve's Steam Deck Steamworks documentation confirms both anti-cheat systems support Proton, each requiring a manual configuration step. For Easy Anti-Cheat's Epic Online Services version, publishers must enable Linux as a client platform in the SDK configuration, activate a Unix platform module, and include the Linux library alongside the Windows one in their depot. BattlEye requires coordination with a Valve or BattlEye technical contact. Neither is technically complex, but both require publisher action, and many publishers of high-profile multiplayer games have not prioritized it.

Valve has acknowledged the problem explicitly in the context of the Steam Machine. Engadget reported Valve's statement that the Steam Machine similarly requires developer participation to enable anti-cheat, but Valve expects the commercial incentives to be stronger for the Machine than for the Deck, given that it anticipates more multiplayer gaming on the new hardware. Valve has expressed hope that the Steam Machine launch will shift publisher calculations on anti-cheat support. Whether that commercial reasoning will move publishers remains to be seen.

The consequence is that the games where Linux compatibility matters least -- single-player AAA titles, indie games, games with smaller multiplayer communities -- tend to work well on Proton. The games where it would matter most for player counts -- heavily played battle royale titles, esports games, live-service shooters -- are frequently blocked specifically because of kernel anti-cheat. This asymmetry is unlikely to resolve quickly without either a platform-level shift in anti-cheat architecture or sustained commercial pressure from the Linux gaming population.

Which situation describes you?
Strong candidate

You want a machine that boots into games, stays out of the way, and never needs driver maintenance. The immutable OS model was built for you. SteamOS on a Steam Deck or Legion Go S, or Bazzite on AMD desktop hardware, will serve this use case well. The trade-off — inability to freely install system packages — will not affect you because you are not trying to use it as a general computer.

The main question to resolve before committing: do any of the games you play frequently use kernel-mode anti-cheat? Check ProtonDB for each title. If the answer is yes, Linux gaming will block you from those titles specifically, and a dual-boot setup may serve you better than a full migration.

Low friction entry

You have a working Windows install and want to explore without commitment. Bazzite is the lowest-friction entry point — it supports AMD, NVIDIA, and Intel, has a clean installer, and the Steam Gaming Mode experience is indistinguishable from a Steam Deck for single-player games. Keep Windows on a separate drive or partition for any titles that need it.

The shader cache cold-start stutter will be the most noticeable friction in the first few sessions with each game. It diminishes. Do not benchmark your first playthrough of a new title and declare the platform worse than Windows — give the cache time to populate.

Proceed carefully

NVIDIA support on SteamOS-derivative distributions is improving but not uniform. Bazzite is the only recommended distribution for NVIDIA desktop hardware — it ships proprietary drivers and has done the engineering work to make Gamescope function with them. SteamOS itself and the immutable HoloISO do not provide NVIDIA support reliably.

Valve's own developer has confirmed that open-source NVIDIA driver integration (NVK) is "still pretty nascent" and is one of the two blocking factors for a general SteamOS desktop release. If you have current-generation NVIDIA hardware and want a SteamOS-like experience today, Bazzite is your path. Expect occasional rough edges, especially with newer NVIDIA architectures and Gamescope's DRM/KMS frame-flip path.

Depends on hardware

Steam Deck and Lenovo Legion Go S: you are already on SteamOS. No action needed, and you already have the most polished gaming Linux experience available. ASUS ROG Ally and original Legion Go: SteamOS 3.7.8 officially supports you with an AMD GPU and NVMe SSD requirement — check the Valve installation docs before proceeding. MSI Claw and Intel-based handhelds: SteamOS does not currently support your hardware. Bazzite is your best option for a Steam Gaming Mode experience.

Handheld-specific features — TDP control, suspend/resume, button mapping, display rotation — are significantly more polished on official SteamOS than on third-party distributions. If official SteamOS support is available for your device, prefer it over Bazzite for the handheld use case specifically.

Practical Guidance: Choosing Your Platform in 2026

For someone reading this in 2026 and deciding how to approach Linux gaming, here is an honest, specific summary of where each option stands — including approaches that rarely appear in general coverage.

HoloISO (immutable): Still available at github.com/HoloISO, built from actual SteamOS components, and the conceptually closest option to running the real thing on non-Deck hardware. AMD GPU is required for a reliable experience. The maintenance team is smaller than Bazzite's and the long-term trajectory is less certain. Best suited to users who specifically want the SteamOS Arch lineage and are running AMD hardware that does not yet have official SteamOS support.

Bazzite: The current community consensus recommendation, and for good reason. It supports AMD, NVIDIA, and Intel GPUs out of the box, ships with proprietary NVIDIA drivers, has a larger active development team, and provides per-device configuration tools through its ujust utility. Fedora Atomic as its base means it behaves differently under the hood from SteamOS in ways that matter when troubleshooting: package management, update mechanisms, and the base system are Fedora, not Arch. For most users on most hardware, this is the correct starting point.

ChimeraOS: Immutable, predates modern SteamOS, and uses GNOME rather than KDE Plasma for its desktop mode. A solid independent-lineage option if GNOME suits your workflow or if you want a project that does not depend on either Valve's or Fedora's roadmap. Less community activity than Bazzite but actively maintained.

SteamOS itself (via Steam Deck, Legion Go S, or official install): The reference implementation. As of SteamOS 3.7.8, officially supported on the Lenovo Legion Go S (ships pre-installed), and installable with documented caveats on the Lenovo Legion Go and ASUS ROG Ally. Requires AMD CPU and GPU hardware with an NVMe SSD. For general AMD desktops, the installer exists but there is no official support path. NVIDIA desktop hardware and Intel-based handhelds remain unsupported until open-source driver work matures.

Gamescope-session on a traditional distribution: An approach that receives less attention but suits power users well. Install Arch, Fedora, or NixOS as a standard mutable Linux system, then install Gamescope and Steam Gaming Mode session alongside it. This gives you the full gaming interface of SteamOS without the read-only root, so you retain traditional package management, background services, and system-level customization. The trade-off is that updates are your responsibility — there is no atomic update mechanism to roll back if something breaks. This path works particularly well on NVIDIA hardware where Bazzite's Gamescope integration still has rough edges, because you can control the exact driver version and session configuration manually. Documented setup guides exist on the ArchWiki and Bazzite community forums.

NixOS with a gaming configuration: A less mainstream but technically compelling option that has matured considerably. NixOS's declarative configuration model means your entire gaming environment — drivers, Proton version, Gamescope session, Steam settings — can be expressed as a reproducible config file, version-controlled, and rolled back if any update breaks anything. The nixpkgs gaming module and community configurations like Jovian-NixOS (a NixOS port of the Steam Deck's session environment) make this more accessible than it was two years ago. The learning curve is steeper than Bazzite, but for users already familiar with NixOS, it offers the same gaming capability with far more control over the system state.

Distrobox as a hybrid layer: Often overlooked as a standalone solution rather than just a workaround. On any immutable gaming OS — SteamOS, Bazzite, or HoloISO — distrobox lets you run a full mutable container with any Linux distribution's package manager available inside it. For users who need a few specific non-Flatpak tools alongside their gaming setup, a persistent distrobox container using Arch or Fedora provides package manager access without compromising the immutable host. The container shares your home directory and network, and applications inside it can appear in the host's application launcher. This is not a general-purpose Linux desktop replacement, but it closes the gap for users whose only objection to immutable systems is the absence of pacman or dnf.

Purchasing a "Powered by SteamOS" device: For users who do not want to configure anything, this is increasingly the most practical path. The Lenovo Legion Go S ships with SteamOS pre-installed and is priced below its Windows equivalent. Valve has confirmed more partner devices are in development. Buying into the official hardware ecosystem means Valve-managed updates, official support documentation, and hardware-specific tuning (TDP profiles, button mapping, suspend/resume) that third-party installs approximate but rarely match precisely.

What Not to Use

The original (mutable) HoloISO is archived and receives no updates. SteamOS 2 (Debian-based) is ancient and incompatible with current Steam client requirements. Do not install either on a system you intend to use as a daily gaming machine.

What HoloISO Actually Proved

HoloISO is worth understanding not just as a piece of software history, but as a proof of concept that had real-world significance. It demonstrated, before Valve was ready to demonstrate it officially, that SteamOS 3 could run on generic x86_64 hardware, that Proton and Gamescope were not fundamentally dependent on Steam Deck-specific silicon, and that the community demand for a SteamOS-like desktop experience was strong enough to sustain a one-person project for two-plus years.

The project also surfaced the real constraints: NVIDIA incompatibility, the need for publisher cooperation on anti-cheat, the tension between immutability and power-user flexibility, and the sustainability challenges of a project run out of a Telegram channel by a student with no organizational backing.

Those constraints are not resolved by HoloISO's successor projects -- they are inherited by them. Bazzite has solved the NVIDIA problem and the sustainability problem. The anti-cheat problem remains structural. The immutability trade-off is now a deliberate design choice across the board, not an accident. And Valve's own SteamOS, when it arrives on desktop hardware officially, will face the same questions HoloISO faced: which GPUs, which anti-cheat solutions, which display server, which hardware configurations get official support and which ones do not.

Linux gaming in 2026 is genuinely better than it has ever been. Proton 10 works. Gamescope works. The Deck Verified catalog has over 21,000 titles. The Steam Deck sold at commercial scale and normalized the platform. SteamOS now runs officially on third-party handhelds for the first time. The Steam Machine is still en route -- delayed by a global RAM shortage but confirmed for 2026 as recently as GDC in March 2026. HoloISO's contribution to that trajectory was real, even if the project itself is now an archived repository with a note that the original is no longer maintained. It filled a gap that needed to be filled, and it filled it long enough for the broader ecosystem to catch up to where HoloISO pointed.

The deeper pattern here is not about HoloISO specifically — it is about how open platforms evolve. Community projects do not just serve users in the short term. They build existence proofs: demonstrations that a thing is possible, that there is demand, that the engineering is tractable. Valve did not create Bazzite. Valve did not create the demand for a SteamOS-like desktop experience. The community created both, and Valve's institutional response followed. Bazzite is not HoloISO's replacement — it is evidence that what HoloISO proved was worth building on, more sustainably, with more people, and with the organizational backing the original project never had. That is how open ecosystems are supposed to work. HoloISO worked exactly as it was supposed to.

Sources