From Rival to Contributor: How Microsoft Went from Calling Linux a Cancer to Building Its Own Distribution

From Rival to Contributor :penguin:

It’s funny how things change over time.

There was a time when Microsoft and Linux were far from being friends. Microsoft saw Linux as a threat to Windows and its business model.

But today, Microsoft is doing something that would have sounded pretty strange back then: contributing to Linux, supporting open-source projects, and even maintaining its own Linux distribution.

So, how did we get here?

  1. “Linux is a cancer” (2001)

Back in 2001, Steve Ballmer, Microsoft’s CEO at the time, famously called Linux a “cancer”.

Microsoft was heavily focused on Windows, while Linux followed a completely different approach. It was open source, people could modify it, share it, and build on top of it.

Naturally, the two sides didn’t get along very well.

  1. Things started getting interesting (2009–2011)

In 2009, Microsoft contributed more than 20,000 lines of code to the Linux kernel, mainly to improve support for running Linux on Hyper-V.

And by 2011, Microsoft had become one of the top five corporate contributors to the Linux kernel.

Quite a change from the “Linux is a cancer” days, isn’t it? :joy:

  1. Microsoft starts changing (2014–2020)

Over the following years, Microsoft became much more involved in the open-source world.

A few important moments:

  • 2015: Microsoft released Visual Studio Code, built on the open-source Code - OSS project.
  • 2018: Microsoft acquired GitHub, one of the biggest platforms used by open-source developers.
  • 2020: Microsoft introduced CBL-Mariner, its own Linux distribution, designed for internal infrastructure and cloud-related workloads.

At this point, it was pretty clear that Linux had become an important part of Microsoft’s plans.

  1. From CBL-Mariner to Azure Linux

CBL-Mariner eventually evolved into what we now know as Azure Linux.

In 2023, Azure Linux became generally available as a container host for Azure Kubernetes Service.

And on November 17, 2026, Microsoft announced the general availability of Azure Linux 4.

It’s a Fedora-derived, RPM-based distribution designed mainly for cloud workloads. Version 4 comes with newer components, including Linux 6.18 LTS and DNF5.

So yes, Microsoft now develops and maintains its own Linux distribution while also contributing to the wider Linux ecosystem.

Imagine explaining that to someone back in 2001. :joy:

  1. What do I think about all this?

Personally, I don’t think Microsoft suddenly fell in love with Linux.

I think they simply realised that fighting Linux wasn’t going to get them very far.

Linux became too important in cloud computing, servers, containers, development, and modern infrastructure. Ignoring it was no longer a realistic option.

And honestly, I don’t see a problem with Microsoft contributing to Linux. That’s what open source is supposed to be about: people and companies working together, even when they have different interests.

Of course, Microsoft has its own business goals. I’m not pretending otherwise.

But if those contributions help improve Linux for everyone, then I think that’s a good thing.

As someone working on my own Linux distribution, Gamma Linux, I find this whole story quite interesting. It shows just how far Linux has come, and how even companies that once saw it as a threat are now building their own products around it.

I’m curious about your thoughts:

Do you think Microsoft genuinely changed its approach to open source, or did it simply adapt because the tech industry changed?

And one more question:

Would Linux be where it is today without contributions from companies like Microsoft, Red Hat, Intel, and others?

Feel free to share your thoughts. Just keep the distro wars out of here. :joy::penguin:

:books: Sources & Further Reading

  1. Microsoft Azure Blog — “Microsoft’s Open-Source Journey: From 20,000 Lines of Linux Code to AI at Global Scale” (https://azure.microsoft.com/en-us/blog/microsofts-open-source-journey-from-20000-lines-of-linux-code-to-ai-at-global-scale/)

    • Microsoft’s Linux kernel contributions in 2009 and its ranking among the top five contributors in 2011.
  2. Microsoft Community Hub — “Azure Linux 4 Is Generally Available” (https://techcommunity.microsoft.com/blog/linuxandopensourceblog/azure-linux-4-is-generally-available/4573381)

    • Official Azure Linux 4 announcement, November 17, 2026.
  3. Microsoft Community Hub — “Built by Experience: The Journey to Azure Linux 4.0” (https://techcommunity.microsoft.com/blog/linuxandopensourceblog/built-by-experience-the-journey-to-azure-linux-4-0/4559114)

    • The evolution from CBL-Mariner to Azure Linux 4.0.
  4. Microsoft Azure Updates — “Azure Linux Support in AKS — General Availability” (https://azure.microsoft.com/en-au/updates?id=generally-available-azure-linux-support-in-aks)

    • Azure Linux became generally available as an AKS container host in May 2023.
  5. Microsoft Learn — “Azure Linux Architecture Overview” (Azure Linux Architecture Overview | Microsoft Learn)

    • Technical architecture, Fedora ecosystem, RPM packaging, and kernel information.

This post combines historical events, Microsoft’s official announcements, and my personal interpretation of its shift toward open source.

5 Likes

I agree with you. I worked at Digital Equipment Corporation from 1985-1997; we went through similar things. When I arrived I worked in the Telecommunications Systems Engineering Group TSE/G, and we directly supported the Telcos: AT&T, the Bell Operating Companies, and the other telecom companies, though AT&T and the regional companies, by far, were the focal point of our work.

Back then, we ported our latest device drivers to UNIX System V for the models we were selling. We also had Ultrix 11, 11M, 32, etc. in the eighties. Then in the 1990s we joined the OSF and worked alongside other hardware vendors to support the Common Desktop Environment. By that time we had built Digital UNIX, using both System V AND BSD libraries.

Before then the President of Digital had gone on record to say that “Unix is snake oil”, but of course that was just his opinion; his ideas led to good network interoperability in the company’s featured operating system, VAX/VMS with DECnet as the networking engine.

However, as Unix became popular the TCP/IP networking stack became common across Unix systems and DECnet meant nothing to other systems; only Digital, with interoperability between their different OS’s took advantage of DECnet, so what was once their cash cow became their noose as other systems lowered those costs considerably.

Moreover, that same executive wondered why anyone would want a personal computer so you can see that as times change, situations change; at least Microsoft woke up to that reality. Digital ended up getting absorbed by Hewlett Packard, who went to battle with IBM from top to bottom. Both were strong at the enterprise systems level, but IBM was the next to make a move:

They determined that while sales of various computer lines were strong, it was VERY CLEAR that only their mid to high end enterprise systems were delivering the margins they were seeking, so they sold their entire personal systems businesses to Lenovo, who has produced the replacements for the IBM Thinkpad line for a couple of decades now.

I think that Microsoft, being a large company, still wants to retain as many of their legacy systems as long as they are still profitable, but they are not stupid; the sales of Linux systems that integrate well with their products and services sustains profit margins.

Three of the large current contributors to Linux are IBM, Hewlett Packard and Microsoft; I believe that Oracle and a few other substantial companies also make various contributions and to me that’s what makes it useful to everyone; I neglected Intel; they have long contributed too, as have several other semiconductor companies, and that’s what has caused the ecosystem to grow so large, but it’s also what makes it function on such a broad range of hardware.

There are pluses and minuses to this and frankly to just about anything, but on the whole this is close to what I envisioned between 1999-2001 as I was writing my graduate school papers at the online University of Phoenix.

At that time I noted that Linux had NOT made much of a dent in the desktop market at all, but I predicted that it would end up being a significant force in the mobile computing environment and this would drive some work back to server environments and possibly even the desktop as a critical mass finally emerged.

Linux didn’t do ALL of that by itself, but a modified Linux kernel IS found in Google network products. Other network growth projects have emerged out of the Berkeley Standard Distribution (BSD). That in itself is an interesting development.

AT&T used to allow sabbatical leaves for either teaching or learning at other institutions. I believe that either Ken Thompson, Dennis Ritchie or possibly both of them had a time where they were at the University of California at Berkeley and that’s where some of what we know as BSD emerged.

However there were “fireworks” taking place between Berkeley and AT&T in the early 1990s, right around the time that Linux and BSD Unix were getting “attention”. The legal battles actually benefited Linux and slowed BSD development and hardware support.

I’ve checked BSD a few times in recent years, hoping to see better results; the last time I tried FreeBSD it DID NOT come directly with wireless capabilities - in any case they were not automatic. Where I live, that’s a show stopper because I do not have ANY wired access; that makes it extremely difficult for me to install a stand alone system where I have to GRAFT in any wireless capabilities, so despite an interest in revisiting one or more BSDs again, ONLY versions with PRE-CONFIGURED wireless support AND fairly recent WiFi card support will work for me.

Anyway, I think that despite the vehemous dislike that some people, particularly European and Asian residents have for anything backed by a US conglomerate, reality is that those places and their employees have contributed a LOT; Torvalds himself has worked for more than one US company; though Finnish in origin, as far as I know he continues to live in the USA; so do many of the leading kernel developers; anyway, to me an international cooperation is what makes all of these efforts work to their best and even competing companies now realize this.

4 Likes

Micro$oft has always been behind Unix and the Posix standard. They finally figured it out that they were.
They (M$) have a history of borrowing ideas to get more money, they “borrowed” Apple’s GUI (originally made by Xerox) when coming up with the word processor software.
They also came up with the design for Windows Server 2008’s two-way transitive trust from Novell’s Netware. So them becoming part of Open Source is something that was expected in my eyes, they finally figured out that Linux was built on Unix and is more advanced.
They also figured out that bad-mouthing their competitors isn’t a great idea either, look how well it worked for Apple/Macintosh (anyone remember the Superbowl commercials that MAC put out bashing Microsoft?).
And you forgot to mention 2025 - them recommending people switch to Linux if they couldn’t afford a computer that met Windows 11’s minimum requirements.
For me I was already there, I just used Windows to play games and for Adobe, so paying $30 USD for the Windows ESU was worth it.

5 Likes

Yeah, Digital did somewhat better when they joined the “movement” instead of arguing about it, but they didn’t have an effective way to remain a profitable entity and did better by selling their assets (calling it a “merger”, but it was really a sell off).

Both Sun Microsystems and Microsoft undercut Digital in the nineties.

Microsoft recovered from their gradual decline by hopping onto the next positive trend; I think that the best leading companies anticipate change and embrace it; if they are really good they can INITIATE change and benefit from it.

I’m not 100% sold on the actual value of “artificial intelligence” or AI and here’s why: the idea has merit, but just slapping a badge on something that’s actually more of a market information grabbing vehicle is borderline criminal behavior, so unless you know both the actual cost and the benefit of the technology you are considering using, DO so BEFORE blindly going into it, and at least then you can decide whether the cost of adopting a technology that “gathers information” about you is worth it. Google has gathered market information about us for a couple of decades and they are billionaires as a result. How much of that do they “give back” in positive ways: (I don’t know the answer), if they do, that’s at least some “compensation”, but it should be a multi billion dollar human investment too, otherwise it’s like sharing crumbs.

Thoughts and counterpoint on these comments?

4 Likes

This is the best remark I read this week about A.I. ! :+1:

5 Likes

Thanks! As I tried to write, it’s not that ALL of it is 100% bad, but for the most part quite a bit of it is marketing hype.

“Machine learning” CAN have SOME value in cases where you want a robot or some mechanical device to reliably REPEAT certain tasks. REPLACING the human mind, at least in my opinion, is NOT the objective, but some power hungry people believe that they can hoist such ideas on an unsuspecting public, and that’s where we can help.

I don’t want to completely wipe out any ideas that advance actual innovation and POSITIVE change, but rather than just change things, I’d rather CHANGE ONLY the things that we can offer that improve either the results, the experience, or both; if it can’t do at least one of those, to me it’s more like those WAR Terminator devices in the Science Fiction movies and that is BAD stuff, aimed at destruction rather than positive change.

So if someone can show me some “positive change” AI efforts, I’ll look and listen, but I for one will be SLOW to adopt them; I’m too old to play “John Connor” in the Terminator and save the world, but I can write with some degree of intelligence about what can be helpful and what can be destructive.

Create genuinely good, quality food and clothing that can be donated from the billions in profits made from software and MAYBE I’ll give such efforts a thumbs up, as long as I don’t also see some asterisk beside the effort where the recipients become servants of such a “machine”!

5 Likes

But the history of computing is littered with companies short shortsightedness, examples include HP’s letting Steve Wozinak’s smaller chips go, Steven Jobs trusting Bill Gates and the list goes on.
As far as AI goes,there’s part of me that wishes it had stayed with NASA rendering simulations of how the universe was created (those were cool to watch) and part of me hopes that governments learn to use it to gather accurate information on the cost of living for their countries.
I think currently companies are using it to replace workers claiming AI can do everything a human can do, there’s a pipe dream if I ever heard one :joy:

5 Likes