SoylentNews is people
SoylentNews
By my count, Linux has over 11% of the desktop market. Here's how I got that number - and why people are making the leap:
My colleague Jack Wallen and I have been telling you for a while now that you should switch from Windows to the Linux desktop. Sounds like some of you have been listening.
The proof of the pudding comes from various sources. First, with Windows 10 nearing the end of its supported life, we told you to consider switching from Windows to Linux Mint or another Windows-like Linux distribution. What do we find now?
Zorin OS, an excellent Linux desktop, reports that its latest release, "Zorin OS 18 has amassed 1 million downloads in just over a month since its release." What makes it especially interesting is that over "78% of these downloads came from Windows" users.
[...] Many have already been making the leap. By May 2025, StatCounter data showed the Linux desktop had grown from a minute 1.5% global desktop share in 2020 to above 4% in 2024, and was at a new American high of above 5% by 2025.
In StatCounter's latest US numbers, which cover through October, Linux shows up as only 3.49%. But if you look closer, "unknown" accounts for 4.21%. Allow me to make an educated guess here: I suspect those unknown desktops are actually running Linux. What else could it be? FreeBSD? Unix? OS/2? Unlikely.
In addition, ChromeOS comes in at 3.67%, which strikes me as much too low. Leaving that aside, ChromeOS is a Linux variant. It just uses the Chrome web browser for its interface rather than KDE Plasma, Cinnamon, or another Linux desktop environment. Put all these together, and you get a Linux desktop market share of 11.37%. Now we're talking.
If you want to look at the broader world of end-user operating systems, including phones and tablets, Linux comes out even better. In the US, where we love our Apple iPhones, Android -- yes, another Linux distro -- boasts 41.71% of the market share, according to StatCounter's latest numbers. Globally, however, Android rules with 72.55% of the market.
[...] Now, of course, StatCounter's numbers, as Ed Bott has pointed out, have their problems. So I also looked at my preferred data source for operating system numbers: the US federal government's Digital Analytics Program (DAP).
This site gives a running count of US government website visits and an analysis. On average, there are 1.6 billion sessions over the last 30 days, with millions of users per day. In short, DAP gives a detailed view of what people use without massaging the data.
DAP gets its raw data from a Google Analytics account. DAP has open-sourced the code, which displays the data on the web, and its data-collection code. You can download its data in JavaScript Object Notation (JSON) format so you can analyze the raw numbers yourself.
By DAP's count, the Linux desktop now has a 5.8% market share. That may not sound impressive, but when I started looking at DAP's numbers a decade ago, the Linux desktop had a mere 0.67% share. We've come a long way.
If you add Chrome OS (1.7%) and Android (15.8%), 23.3% of all people accessing the US government's websites are Linux users. The Linux kernel's user-facing footprint is much larger than the "desktop Linux" label suggests.
[...] But wait, there's more data. According to Lansweeper, an IT asset discovery and inventory company, in its analysis of over 15 million identified consumer desktop operating systems, Linux desktops currently account for just over 6% of PC market share.
Earlier this year, I identified five drivers for people switching from Windows to Linux. These are: Microsoft's shift of focus from Windows as a product to Microsoft 365 and cloud services, the increased viability of gaming via Steam and Proton, drastically improved ease of use in mainstream distros, broader hardware support, and rising concern about privacy and data control.
Three others have emerged since then. One is that many companies and users still have perfectly good Windows 10 machines that can't "upgrade" to Windows 11. ControlUp, a company that would love to help you move to Windows 11, has found that about 25% of consumer and business Windows 10 PCs can't be moved to Windows 11.
[...] Another is that many people really, really don't want to move to Windows 11. A UK survey by consumer group Which? in September 2025 found that 26% of respondents intended to keep using Windows 10 even after updates stopped. Interestingly, 6% plan to go to an alternative operating system such as Linux.
[...] Finally, not everyone is thrilled with Windows 11 being turned into an AI-agentic operating system. Despite all the AI hype, some people don't want AI second-guessing their every move or reporting on their work to Microsoft.
After Microsoft president Pavan Davulur tweeted on Nov. 10 that "Windows is evolving into an agentic OS, connecting devices, cloud, and AI to unlock intelligent productivity and secure work anywhere," he probably expected Windows users to be happy with this vision. They weren't.
[...] My last reason for people looking to Linux from Windows doesn't matter much to users in the US, but it matters a lot to people outside the US. You see, the European Union (EU) governments don't trust Microsoft to deliver on its service promises under potential US political pressure.
This has resulted in the rise of Digital Sovereignty initiatives, where EU companies and not American tech giants are seen as much more trustworthy. As a result, many EU states have dropped Microsoft programs and have switched to open-source software.
That includes the desktop. Indeed, one EU group has created EU OS. This is a proof-of-concept Linux desktop for a Fedora-based distro that uses the KDE Plasma desktop environment.
It's not just the EU. The UK also no longer trusts Microsoft with its data. A 2024 Computer Weekly report revealed that Microsoft told Scottish police it could not guarantee that data in Microsoft 365 and Azure would remain in the UK.
[...] Taken together, all these shifts make Linux less of a tinker's special and more of a pragmatic option for people who want out of the Windows upgrade treadmill or subscription model.
Desktop Linux is moving from perennial underdog to a small but meaningful slice of everyday computing, especially among technically inclined users, non-American public-sector agencies, and ordinary consumer and business users who want a cheaper, more trustworthy desktop.
(Score: 0) by Anonymous Coward on Sunday December 07, @10:10PM (9 children)
Exactly... I have no idea why this was modded flamebait.
The fact that there is no unified app base for Linux the way there is for macOS and Windows is a HUGE deal for third party BINARY app developers.
(Score: 3, Informative) by Thexalon on Sunday December 07, @11:33PM (2 children)
There was a real attempt at fixing that: The Linux Standard Base (LSB) [linuxfoundation.org]. It didn't catch on because Debian and Red Hat refused to play nice with each other.
"Think of how stupid the average person is. Then realize half of 'em are stupider than that." - George Carlin
(Score: 1) by Bentonite on Monday December 08, @09:43AM (1 child)
>Linux Standard Base
>Look inside; https://refspecs.linuxfoundation.org/LSB_5.0.0/LSB-Common/LSB-Common.pdf [linuxfoundation.org]
>GNU library, GNU library, GNU library, GNU library, GNU library, Mozilla library, Mozilla library, PAM, GNU library, GNU library, openssl, GNU library, GNU library and zlib.
>Bunch of OpenGL libraries, Qt, Xorg libraries, freetype, cario, GNOME, GtK, cups etc - you get the idea.
That clearly should have been called the "GNU Standard Base (GSB)", as well it doesn't even define what minimum version number of Linux is required (spoiler; as the version of Linux is mostly irrelevant - what's relevant is the GNU libraries and all the other libraries that depend on such).
Most distros seem to follow the GSB, or allows the user to install all the libraries as needed - as even Gentoo seems to have many of the libraries installed by default and packages freetype and cairo and Qt and GtK etc.
(Score: 3, Touché) by Thexalon on Monday December 08, @01:01PM
Right - the point of LSB was trying to create a binary compatibility option, which hasn't really happened. Instead, we get source compatibility where most stuff will run on most Linuxes after a "./configure && make && make install", and that means that distros can easily package a lot of it, which is still good and useful, but not really doing the job for people trying to put proprietary software on top of a Linux desktop.
Of course, there's arguments to be had in multiple directions about whether putting proprietary software on top of a Free Software platform is a good goal or not.
"Think of how stupid the average person is. Then realize half of 'em are stupider than that." - George Carlin
(Score: 1) by Bentonite on Monday December 08, @09:27AM (5 children)
How proprietary software developers have trouble offering proprietary software that takes the users freedom for GNU/Linux is a feature, not a bug.
If Linux's unified singular SYSCALL interface was targeted only, such binary will in fact work with any variant of GNU/Linux (no matter how many parts of GNU have been swapped out for something worse) - but that interface is really limited.
(Score: 0) by Anonymous Coward on Monday December 08, @09:21PM (3 children)
Linus considered it a bug: https://www.youtube.com/watch?v=Pzl1B7nB9Kc [youtube.com]
But I mean what does he know, he's just Linus. Whereas Bentonite is Bentonite.
(Score: 1) by Bentonite on Tuesday December 09, @01:03AM
Linus has stated many times that he does not care about the users freedom - he only cares about technical things, like how functionally good software can immediately be and how many users it has - thus he has permitted Linux to contain proprietary software, encouraged people to run proprietary software countless times and he has developed several other proprietary programs.
Of course such developer would consider that maybe if more proprietary software (that takes the users freedom) was available and developed for GNU/Linux, that might indirectly increase the popularity of his kernel and therefore regards anything preventing the popularity of such proprietary software as a bug.
He refuses to even consider that following a faster road, that goes to the wrong place is a bad idea (as clearly you end up in the wrong place and you stay there forever, or have to trek back to the right road, which ends up slower than just following the road to the right place even though such road is slower).
Btw, please don't recommend that people run proprietary software by linking to youtube.com - if for some reason a video is on youtube, in the past generally could could do; `yt-dlp https://www.youtube.com/watch?v=Pzl1B7nB9Kc` [youtube.com] - too bad yt-dlp now automatically executes all of the proprietary software sent unless you use an old version (or configure the new version and hope that a codepath is there doesn't get hit).
(Score: 2) by aafcac on Tuesday December 09, @01:48AM (1 child)
I think the zealotry is less than helpful. It's great to have OSS, but if the ability of people to make money on the system with proprietary software or the software can be used in proprietary software is unacceptable, then I have no idea how you ever get enough normies on board to ever reach the year of Linux.
A bunch of the concerns have proven over time to be overblown and things like ZFS are really hard to get working because of barely enforceable licensing terms.And I do mean that, if you don't have the money to hire attorneys, then the difference between GPL and something more permissive is theoretical.
(Score: 1) by Bentonite on Tuesday December 09, @08:49AM
>I think the zealotry is less than helpful.
Only zealots get the freedom done - thus zealotry is extremely helpful.
>It's great to have OSS
It is not great to have "OSS" - "open source" was announced in 1998 as an attack on free software; http://catb.org/~esr/open-source.html [catb.org]
>but if the ability of people to make money on the system with proprietary software
Proprietary software is totally unacceptable no matter what, as it gives the developer (and the developers employers) unjust power over the users; https://www.gnu.org/proprietary/ [gnu.org]
Anyone sane would regard proprietary food recipes as unacceptable no matter what (too many people used recipes to make them proprietary), but somehow you are insane if you regard proprietary software to be unacceptable no matter what (software was successfully made automatically proprietary without a free license, as not enough people used software at the time).
There are plenty of legitimate ways to make money on the system that respects the users freedom - via support, warranty, training, custom development, hosting and much, much more.
Only an extremely limited number of companies and individuals are able to profit from proprietary software, as existing masters do absolutely everything to prevent competition.
It seems there's as just as much money, or more money in custom software (which can trivially be free software if the customer has the purely financial wit to make the contract require that source code is provided and that the software may not be programmed to depend on proprietary dependencies).
>or the software can be used in proprietary software is unacceptable
Weak licenses are a sad case, as while the developer provides the software under a free license (or maybe not if the developer is a patent troll, as many weak licenses don't have a patent license), if you go by the primary usage of such software, it's often the case that it has been incorporated into proprietary software and therefore the software is not free for most of its users.
Clearly nobody shouldn't write proprietary software totally gratis - they should use a strong license instead, or if they are going to write proprietary software, at least they should ensure to get paid.
>I have no idea how you ever get enough normies on board to ever reach the year of Linux.
The year of GNU/Linux was 1995 - as finally you could use a computer in freedom again (alas that freedom was savagely ripped away with the addition of the first of many proprietary programs into Linux in 1996).
Popularity of the kernel, Linux has already been achieved - every computer that matters runs GNU/Linux and even more than half of normies use that kernel (used in Android as proprietary software, as the user almost always doesn't get the source code and can't change it).
I don't consider that even more normies soiling GNU with more proprietary software, never realizing that the OS was developed so the user could have freedom, would be a good thing.
>things like ZFS are really hard to get working because of barely enforceable licensing terms
GPLv2 and CDDLv1.1 are incompatible licenses, end of story.
"openZFS" has only survived as the Linux developers don't enforce their license.
The "openZFS" problem would be quite easily solved if the Linux developers enforced their license and demanded that it be re-licensed to a compatible license and then it could be mainlined, but for some reason they won't do that?
>if you don't have the money to hire attorneys, then the difference between GPL and something more permissive is theoretical.
You don't need an attorney to advise that a license will be permanently terminated unless it is complied with.
You'll likely easily find a no-win-no-fee attorney happy to accept such a easy copyright case - either the company has a license to distribute the software, or is committing copyright infringement for profit.
(Score: 2) by mcgrew on Tuesday December 09, @08:41PM
How proprietary software developers have trouble offering proprietary software that takes the users freedom for GNU/Linux is a feature, not a bug.
Indeed, although I'd call it a design defect. Libre Office, GIMP, Audacity, Handbrake, all have no problem writing programs that run under damned near any OS (although Windows versions of Handbrake and Audacity are pretty much crippled in Windows).
The same small group of people own all of the world's big businesses.
Why do the mainstream media act as if Donald Trump isn't a pathological liar with dozens of felony fraud convictions?