I don't actually believe Linux desktop user marketshare is what's being recorded in their metrics. It's more likely local AI, possibly home users using Linux computers that run local AI, but it's web scrapes not what's typically defined as home user activity. There's no real way to measure the desktop Linux market because there isn't a centralized licensing tracker except for the small percentage using paid options from Redhat, SuSE, or Canonical.

I don't think it's critical to knowing of Linux is more relevant than it was 10 years ago. It certainly is, and we know anecdotally that many more people are using it outside of servers and a small number of corporate workstations. The rise of local AI and Steam OS (which has a Linux desktop mode) is causing most of it