Linux kernel maintainers are removing support for the Intel 486 processor - a chip that debuted in 1989. Linus Torvalds himself says there's "zero real reason" to keep supporting it.
This is the end of an era for legacy computing.
Linux has always prided itself on running on anything. Ancient hardware, embedded systems, obscure architectures - if it can execute code, someone has probably gotten Linux running on it. That philosophy made Linux dominant in servers, supercomputers, Android phones, and basically everything that isn't a Windows PC or Mac.
But even Linux has limits. And supporting the i486 has finally crossed the line from "useful flexibility" to "pointless burden."
The i486, for those who don't remember, was Intel's fourth-generation x86 processor. It powered early 90s PCs, introduced pipelining and integrated floating-point units, and was genuinely impressive technology in its time. But it's 2026 now. That chip is 37 years old.
Linus Torvalds, who created Linux and still maintains the kernel, was clear in his reasoning: there's no one left using i486 systems in any meaningful way. The code to support it adds complexity, makes other improvements harder, and provides effectively zero benefit to anyone.
This isn't about planned obsolescence or forcing people to buy new hardware. If you want to run Linux on a 486, you can use an older kernel version. It'll still work. But new kernel development won't accommodate that ancient architecture anymore.
What's interesting is what this tells us about the costs of backwards compatibility. Every time you maintain support for old systems, you're making tradeoffs: more complex code, more testing, more edge cases, more things that can break. At some point, the costs outweigh the benefits.
Microsoft faces this constantly with Windows, which is why they still support hardware from the mid-2000s but have drawn lines about what older systems can run newer versions. Apple is more aggressive about dropping old hardware, which is part of why their software updates tend to work more smoothly - fewer edge cases to handle.
Linux, by culture and necessity, tends to support things longer. Server hardware from 2010 is still in production use. Embedded systems built in 2005 are still out there doing their jobs. Linux's ability to run on that hardware is part of its value.
But i486? That's not a server running in a datacenter. That's not an embedded system in industrial equipment. That's someone's hobby project or a museum piece.
The decision to drop i486 support is also a signal about where Linux development is heading. Modern features like better security, more efficient power management, and support for new hardware all assume certain baseline capabilities. Carrying ancient architectures along makes everything harder.
For most Linux users, this won't matter at all. If you're running Linux on a laptop, server, or Raspberry Pi, nothing changes. But for the broader philosophy of open-source software, it's a reminder that "runs everywhere" has practical limits.
The technology was impressive in 1989. The question is whether supporting it in 2026 makes sense. Linus Torvalds says no, and he's probably right.
