No, Linux's share of the desktop market in North America hasn't exceeded 10%, and Windows hasn't suddenly started losing millions of users. 2026 won't be the year of Linux, for those who care. On the contrary, it might be the year the internet collapses, as the sudden surge in numbers appears to be entirely driven by AI bots.

Almost every month, a viral story emerges claiming that Windows is losing market share and Linux is rapidly gaining ground. These claims invariably cite the same source: StatCounter, whose data is widely misinterpreted.

For example, this week, one of the most-voted posts in the Linux community on Reddit and even on Hacker News claimed that "Linux's share of the desktop market has surpassed 10% in North America." The original poster included a huge image showing Linux's triumph after years of struggle, but it also highlights the obvious problem with the numbers:

the data shows Windows at 57.54%, ChromeOS at 2.06%, Linux at 10.65%, OS X at 21.14%, and macOS at 8.6%. Here I stopped reading. We're looking at OS X, which was superseded by macOS in 2016, at 21.14%, while macOS itself is at 8.6%.

Even if we consider OS X and macOS as separate systems, it doesn't make sense that the older OS X has more users than macOS.

StatCounter doesn't actually track Windows or Linux users

, but is the problem really StatCounter's? Actually, no. It was never meant to be used as a primary benchmark for market share on desktops or mobile devices. For those unfamiliar, StatCounter is an analytics platform for web developers and administrators, allowing website owners to monitor their traffic if they prefer not to use Google Analytics or other options.

Unlike Google Analytics, StatCounter regularly publishes public reports on "market share," with final figures typically appearing at the end of each month, and it also publishes its methodology.

According to StatCounter: "Our statistics are based on over 3 billion monthly pageviews from over one million websites globally, and we are not aware of any other public service that provides market share statistics with a larger sample size. We base our statistics on pageviews (not unique visitors) and believe this gives a fairer estimate of internet usage."

As you can see, StatCounter counts pageviews, not actual people or computers. It cannot and will not collect such sensitive data. Only Microsoft or its manufacturing partners can accurately tell you how many Windows users there are now compared to five years ago.

In other words, if Linux-based agents or bots suddenly visit sites using StatCounter's tracking code, Linux numbers might rise, and Windows' share might fall because the numbers are measured on a 100% scale. But this doesn't mean people have replaced Windows on their machines with Linux.

StatCounter has reported similar anomalies before.

StatCounter incorrectly reported a surge in Windows 7 usage (later corrected; old image).

It's true that Windows 11 isn't a perfect operating system, and some prefer Windows 7 or even Windows 98, but that doesn't mean it's losing users to a much older OS.

In previous reports, Windows 8.1 also unexpectedly surged in some regions, and this was later corrected.