There two major things in my point of view.

First, in my experience, software companies mostly provided developers with Linux/Mac machines (rather the latter) — and they were not only startups. At the same time, the situaton was opposite for companies for which software development wasn't the main focus (almost 100% of them used Windows). I find it reasonable, as software professionals know that Windows suites developments needs purely (I don't want to start a fire but that's what it is, just the experience) and those who are not into software dev directly don't consider it something important.

The numbers refer to the same PC, meaning the developer bought it, so it is a personal PC, but they also use it for freelance work.

That's also quite true IMO, on a personal level.

Second, the SO survey is hugely overrated (as any dev survey basically) and blindly trusted. The community is too narrow and too opinionated to take the results as something to make conclusions from. Some of them (I neither participated nor read it directly this year tbh, it's too lame, only through the lens of community blog posts focused on some specific aspects, like programming languages) are really quirky.

All in all, I see the results of the OS comparison here as not very meaningful (even though they actually match my own experience, but I know it's quite one-sided).

We're 150 year old business, we provide Macs and Linux machines to our developers, our QA and anyone else who wants them.

Personally I'd use a Mac over pretty much anything as we run everything these days on AWS Linux and so Windows tends to be a bit annoying to context switch (though GitBash etc are mostly used by our Windows users). I'd imagine it's the cost effectiveness of Linux remote instances that is driving the uptake for others.

I'm a freelancer, I use a Linux laptop, and it's the same machine for mostly everything. As an employee that I used to be, I used Windows, then Windows again in another company, then switched to Linux, as I learned our office admins did not support, but tolerated, dual Linux setups. Docker and git were much faster on Linux.

Yet another company had ditched all windows machines before I joined, so I had a Mac mini at work, running MacOS + Parallels so that we could use our inhouse Windows software, at the same time I had another Apple machine, an old Macbook discarded by my former employer, as a personal laptop. When I came back to my former company, I requested a Macbook as a work laptop because I found it was faster and had superior hardware while weighing less than the equivalent Lenovo Windows/Linux alternative. Weight was very important as I used to travel a lot by train.

I still have, and occasionally use, the discarded Macbook and it has served as backup machine for my family, friends, and business partners. I also have another backup/testing machine, an old but also still working Lenovo laptop with both Windows and Linux, just in case.

If I remember correctly, I must have stated that I use Linux for work and leisure, but you can see, that the reality is a little more complicated.

I wonder what effect the recent Crowdstrike outage has on Linux adoption.

My question would be how in the world the numbers are the exact same.

There is a low chance of 31.8% being the same number for personal and pro for Mac, but it's even stranger that Ubuntu would be the same way.