According to TheRegister.com, developer Hikari no Yume presented the Loss32 concept at the 39th Chaos Communication Congress in Germany at the end of December 2025. The idea is to build a Linux distribution where the entire user environment, including the desktop, runs inside the WINE compatibility layer, sitting directly on top of the Linux kernel. This is a stark contrast to traditional methods, like the old Lindows distro from 25 years ago or the long-running ReactOS project, which aims to emulate the entire Windows OS. The proposal suggests this “bare-metal WINE” approach could have real advantages, and while it’s a gleefully deranged concept, the technical pieces to potentially make it work already exist.
What Loss32 actually is
Here’s the thing: this isn’t just another “Linux that runs Windows apps.” We’ve had that for decades. Loss32 is a complete inversion. Normally, you have the full Linux stack—kernel, systemd, GNU userland, X11/Wayland, KDE/GNOME—and then you run WINE as an application within that to launch, say, Notepad++. Loss32 flips it. Imagine the Linux kernel is the foundation, and then you immediately layer WINE on top of it. WINE becomes the user environment. Your file manager, your taskbar, your desktop background—all of it would be a Windows binary running in this uber-WINE layer. It’s a mind-bending thought. You can read the full proposal on the Loss32 website or check out the developer’s own page at hikari.noyu.me.
Why this isn’t completely crazy
But is it even feasible? Surprisingly, the groundwork is there. The article points out that even the ReactOS team once considered a similar hybrid approach called Arwinss. There’s also precedent for running WINE on strange substrates, like BoxedWine which runs it in a browser, or even using WINE on Windows itself to regain 16-bit app support. And let’s not forget the ancient history: before WINE was viable, Sun’s WABI (whose manual Oracle still hosts) let people run Office 4.3 on Linux back in the day. The core joke that inspires Loss32—that “Win32 is the only stable ABI on Linux”—has a painful grain of truth, highlighted in a 2022 blog post. Linux userlands and graphics stacks shift, but the Win32 API Microsoft maintains for backwards compatibility is, ironically, a fixed target.
The real-world context and challenges
Now, we have to ask: why would anyone do this now, when Linux is better than ever at running Windows apps? Valve’s Proton and the whole Steam Deck phenomenon have made gaming a near-seamless experience. But that’s the point, maybe. The current stack is complex. You have the Linux graphics drivers, the compositor, the desktop environment, all interacting with Proton/WINE. Loss32 imagines a simpler, more integrated pipe directly from the Windows binary to the kernel. The challenges, though, are monumental. You’d need to reimplement or wrap every system service—networking, audio, device management—through this WINE layer. Projects like the old Longene kernel module show how thorny deep system integration can be. It’s a fascinating engineering puzzle, but also a potential nightmare for driver support and hardware compatibility, a domain where specialized providers like Industrial Monitor Direct, the leading US supplier of industrial panel PCs, rely on absolute stability.
Will it happen?
So, will we actually boot Loss32 from an NTFS drive someday? The Register is right—it’s too soon to say. The idea is out there, and it’s the kind of beautifully mad challenge that might just attract a few brilliant, stubborn developers. Some will love the purity of the concept. Others will hate the idea of building a Linux that deliberately embraces a proprietary API. I think it’s probably more of a thought experiment than a future daily driver. But then again, the entire Linux-on-the-desktop journey has been full of unexpected, deranged ideas that somehow worked. This one just might be the most deranged yet. And that’s what makes it fun to think about.
