My prediction: Linux market share on Steam hits 10% by early 2028

to be honest hardware prices as they are now is probably a good reason for users to jump ship. Windows 11 is just too heavy on hardware and with users not being able to do hardware upgrades more will be pushed to make the jump instead of spending money they can't afford.

[-]

AnonomousWolf@reddit

It's bad timing sure, but it's still a option many will go for.

Steam deck sold ~2 million units I think. I could see the seam machine at least matching that over 2/3 years

[-]

AlternativePaint6@reddit

The argument for exponential growth is that the more people switch to Linux, the more people will hear about it, the more people will have support around them, and the lower the barrier of entry will be for others as well.

Not saying that will happen or that OP's model is right, but that's an argument to be made.

[-]

flagos@reddit

Well first let's have good support of hardware.

[-]

burning_iceman@reddit

Done. What now?

[-]

flagos@reddit

Lol, my webcam released 3 years ago is still in grey scale.

[-]

burning_iceman@reddit

So? No OS has perfect hardware support. Linux has excellent hardware support in general. The fact that you have one piece of hardware that doesn't work properly is an irrelevant anecdote.

[-]

internet-weirod@reddit

More users demanding better support means that hardware manufacturers will provide it, and Linux works on most hardware fairly well

[-]

ButtonExposure@reddit

This effect is called "critical mass": https://en.wikipedia.org/wiki/Critical_mass_(sociodynamics)

[-]

CORUSC4TE@reddit

I dont think you are wrong, but i dont think there is too much to learn.. Especially if you only want to game and browse.

[-]

scandii@reddit

I feel you're equating inertia with value proposition - there simply isn't a value proposition to switch to Linux for many people.

however we've seen de facto monopolies crash and burn a lot of times throughout history due to the value proposition changing e.g. when was the last time you saw a dvd/blu-ray player or a TomTom?

Linux's value proposition over Windows is mainly there for people that like doing things their way, the fact that it doesn't have a price unlike Windows and that it is free of telemetry.

if these things aren't selling points for you well, why swap.

[-]

bawng@reddit

I think you're also missing how much people hate recent version of Windows. Even my 80 year old dad started asking me for alternatives because Windows is so annoying these days.

[-]

NDCyber@reddit

There kinda is, it is free to use, works well and you won't need to make yourself reliant on one company. Those are at least the points I have heard from a company in the anti discrimination sector when they discussed switching to Linux

[-]

scandii@reddit

the problem is that you just listed three arguments that cancel out. like I am a Linux user, but you didn't sell me one bit on Linux just now.

Go with Ubuntu, now you swapped "reliance on one company" to "reliance on one company". one can argue that the Linux kernel having the Linux Foundation behind it is better than whatever Microsoft are doing, but in reality you're using the OS, not only the kernel.
Windows works well too and has flaws just like Linux distros. I don't vehemently hate the corporate forced AI push but we can't just turn a blind eye that for every issue with Windows we can list one on Linux, and vice versa.
As a company I can see negotiating to not include Windows licenses, but as a private user Windows comes included with most retailers making the cost argument moot.

like, the value proposition just isn't there realistically speaking.

[-]

inemsn@reddit

Go with Ubuntu, now you swapped "reliance on one company" to "reliance on one company"

That is not true. Ubuntu is made and maintained by Canonical but it is by no means entirely reliant on Canonical. If Canonical fucking died tomorrow there would, doubtless, be an immediate gathering of devs and maintainers to form a community that takes up the mantle of maintining Ubuntu, and that's actually possible with Linux distros because they are open source.

but you can't disregard that the caveat of "yeah well if you have this hardware you're forced down this path" is not a concern on Windows.

Correction: It wasn't a concern on windows... until windows 11 rolled around. Now even windows has hardware troubles, and if your PC is a bit too old or a bit too weak you just can't use it.

but we can't just turn a blind eye that for every issue with Windows we can list one on Linux, and vice versa.

I don't understand how this is supposed to be a point against the value in switching to linux.

but as a private user Windows comes included with most retailers making the cost argument moot.

"Free to use" in this context isn't free as in no cost, it's free as in freedom. You have to do some pretty non-trivial hacks just to use windows 11 without tying yourself to Microsoft accounts, for example.

but I also completely see that none of this matters for a person that just wants to run software.

Libre software hasn't mattered for a person that just wants to run software, but increasingly, it does: The freedom linux provides has started to become important to people, especially people outside the US who view reliance on US technology and companies with increasing distrust. People are becoming more and more aware of the dangers of letting the modern digital infrastructure their lives depend on be monopolized by a handful of tech giants and they're starting to take action.

Obviously, this is a slow process: For now, most people still only see this as a rising concern, not as a crisis they need to act on. But it's happening. You can tell that it's happening. Windows is under genuine threat for the first time in forever, because as Linux becomes more and more trivial to install and use, like with Valve's SteamOS, people are more willing to give just a little attention to the arguments in favor of Linux.

[-]

NDCyber@reddit

There is no argument cancelling out anything

Even on Ubuntu you can make changes that Canonical might not agree with and you can also choose another distro if Ubuntu doesn't work. Means you aren't reliant on one company, as you can choose which product of which company you use
They no longer have access to Microsoft Office, after they changed the pricing, and they can't afford to spend more than they do on windows licensing and the now included Copilot. It is also an office, they won't use HDR or Niri, they would probably use something like Mint or debian to write documents. And lets not forget how many faulty updates windows had this year alone, while on Linux you basically hear that everything is normal, besides one security problem
Yeah shit Linux isn't perfect, nobody said that. But it is starting to be better to work with and use
Partially yes, partially no. I talked with my best friend once about laptops and i saw that some Linux laptops were cheaper than the windows alternative, so I told her to maybe look at that next time and just put windows on it, if she wants. She said she would just use what comes with the device, and I think that is true for most people

[-]

SpiritualWillow2937@reddit

Linux is increasingly outperforming Windows in gaming on AMD hardware due to Steam's work on Proton and better AMD support. And of course it destroys Windows on the low end, so who knows, financial hardship might actually boost Linux.

[-]

scandii@reddit

and that's great, but aren't like 95% of all consumer GPU:s Nvidia making this a very niche point to bring up as well?

[-]

malayis@reddit

Consider something like Twitter. After everything that has happened it's still the leading social media platform (at least among its sub-type) and its attempted competitors largely aren't doing great. (and the same could be said for Chrome vs Firefox, Google vs DDG etc.)

Given that's the case, what does that tell us about Linux's actual chances?

[-]

Thevsamovies@reddit

For the longest time the word has been "most games on steam don't work on Linux." That's likely what's holding a lot of people back, who otherwise would be willing to switch.

[-]

aeiedamo@reddit

Ofc but if Linux keeps getting more popular, the average person will be interested in buying a laptop with Linux pre-installed. Framework has already proven this, albeit the average Framework laptop user is more technical than the average Joe, but still, my point stands.

[-]

LuminanceGayming@reddit

linux is on track to have 7.5 trillion percent market share by 2100

[-]

zeth0s@reddit

Seeing that curve, we are at infinite users much earlier. Top class modelling here

[-]

ButtonExposure@reddit

The speed of light (or more precisely the speed of causality) makes it theoretically impossible for Linux to reach an infinite numbers of users. Because even if the universe is infinite, information can propagate through the universe only at a fixed speed. Distributing Linux at a fixed speed in an infinite universe there will always be parts of the universe to which Linux has not yet been distributed. To be able to distribute Linux to infinite users, you'd need an infinite speed of causality.

[-]

stestagg@reddit

you just need to enable the post-einstein scheduler to allow FTL expansion

[-]

Albos_Mum@reddit

The solution to this is clearly to start running the whole universe simulator on Linux, as far as I know it's currently based on TempleOS.

[-]

Moral4postel@reddit

At least that also applies to Windows

[-]

ButtonExposure@reddit

The finite speed of causality would also prevent every computer in the entire universe to be borked from a bad Windows update.

[-]

m1llie@reddit

Yeah, OP needs to fit a logistic curve here, not an exponential. And fitting a logistic curve usually requires having some reliable estimate of the upper limit value, which we don't.

[-]

jaakhaamer@reddit

While the methodology in the post is questionable as others have pointed out, I wouldn't at all be surprised if this actually happens.

We will probably have the Steam Machine and Steam Frame by then (but it is Valve time, so you never know...), which will likely lead to a significant increase just like the Steam Deck did.

Maybe even more, since the console market is orders of magnitude bigger than the handheld market.

[-]

privinci@reddit

This is like that bullshit bitcoin rainbow chart

[-]

AiwendilH@reddit

I suggest a new standard for evaluating technical ideas/suggests/predictions...the NRX factor. Number of relevant xkcd comics:

https://xkcd.com/605/

https://xkcd.com/2048/

https://xkcd.com/2582/

https://xkcd.com/2936/

(And anyone who responds with #927 to this clearly understood the assignment ,))

[-]

redballooon@reddit

Why chose exponential?

Did you compare the mean deviation of the exponential regression with a mean deviation from a linear regression?

[-]

wasabiwarnut@reddit (OP)

Just eyeballing it. In the first few years there's a slowly linear trend which I fitted first and removed from the rest of the data. What you have left is visually a non-linear, increasing growing trend. At an early stage of adoption an exponential behaviour is a fair assumption which, of course, does not necessarily hold up until the predicted year.

[-]

theng@reddit

An exponential here has no meaning: it is impossible to have more user that humans alive (even if we count virtual machines at some point)

https://en.wikipedia.org/wiki/Technology_adoption_life_cycle

with different views on the problem

[-]

redballooon@reddit

Most likely it's a logistic growth curve but it's not clear yet where the upper limit is.

[-]

wasabiwarnut@reddit (OP)

Exactly and, since we don't see any deceleration yet, it doesn't make sense to fit one to the data at this time. It would just introduce another degree of freedom that we can't use.

[-]

ckdx_@reddit

Why did you choose an exponential model? What does the projection look like when a linear trend line is fitted?

[-]

HunsterMonter@reddit

GamingOnLinux has a similar graph and you can clearly see that a linear model is completely inadequate.

[-]

ckdx_@reddit

Do you think the market share will follow the exponential model? Will we hit 12% market share in April 2028?

[-]

Forsaken_Mood9939@reddit

Is this Polymarket ?

[-]

wasabiwarnut@reddit (OP)

No, this is ~~Patrick~~ LibreOffice Calc

[-]

Leverquin@reddit

We should ship Linux kernel into space.

[-]

Great-TeacherOnizuka@reddit

I thought it already surpassed 5% on steam?

[-]

rebootyourbrainstem@reddit

The graph has two data points above 5%, what is your question?

[-]

Great-TeacherOnizuka@reddit

This part:

If the trend holds (hopefully) then we‘d be firmly above 5% by the end of 2026

[-]

rebootyourbrainstem@reddit

Ah, I see. Well, I guess "firmly" is the key word there, as it fluctuates a bit, and last few months have seen especially large jumps.

[-]

wasabiwarnut@reddit (OP)

Momentarily yes but the last two months have been abnormally high. I think it's a statistical fluke and will come down a bit later on.

[-]

rebootyourbrainstem@reddit

The graph has two points above 5%, what is your question

[-]

Unrented_Exorcist@reddit

I like that curve. But as a scientist I have to ask for the error of the model and used equations.

[-]

Lembot-0004@reddit

Yes, correct: prediction of things in the past is much easier.

[-]

my-name-is-puddles@reddit

No, actually, it's easier to predict something in the future because pre-dicting something that is in the present or past is impossible.

[-]

WhJJackWhite@reddit

Isn't that post-dicting?

[-]

acewing905@reddit

Try extending this graph a few more years

[-]

Woodpecker-Visible@reddit

Lets see how microsoft's K2 plan goes of windows debloating.

[-]

Kromieus@reddit

Assuming the singularity doesn’t get us all, i can personally guarantee in 2028 linux will have a minimum of 0% + 1 user (my steam deck)

[-]

BubsyFanboy@reddit

Happens to be after the first 1,5 year of Windows 10's ESU.

[-]

Xotchkass@reddit

My hobby: extrapolating

[-]

Fuckspez42@reddit

With Microsoft both alienating the PC market with their AI crap, and XBox effectively exiting the current-gan console market, this prediction doesn’t sound crazy at all.

[-]

Zatujit@reddit

Disco trends

[-]

Zeikos@reddit

These kind of trends are usually better modelled with S curves

[-]

rebootyourbrainstem@reddit

If these trends continue... ayyy!

[-]

aisingiorix@reddit

Uh, your fish are dead.

[-]

Latlanc@reddit

Holy Cramer of Linux stock. It will go down to 4% instead.

[-]

BashfulMelon@reddit

Buy puts on LNUX if you're so sure about it

[-]

Xatraxalian@reddit

The one thing Steam is possibly missing is that there are also lots of people that play games on Linux without ever touching Steam, like me. I love what Valve is doing for Linux, but I won't buy games there. I've been a GOG.com customer from (almost) the first hour. I like the fact that I can buy a game, download it, save the installation files, and play it for 30 years if I want to, as long as I have a computer to run it on.

(And yes; I still play the original Baldur's Gate from 1998. In 2028, I will be playing it for 30 years. If I ever replay it, I'll play a modded version of the remake from 2012 which removes the idiotic new characters. I also play Caesar, but in its Augustus re-incarnation.)

[-]

Dalnore@reddit

Define "lots". Steam has 130 million monthly users. There's no statistics for GOG, but all the estimates I can find are millions at most. So it might as well be the case that the number of Steam users on Linux (up to 5% in recent surveys, so 6.5 million) might as well be bigger than the total number of people using GOG on any platform whatsoever.

[-]

Xatraxalian@reddit

It is unkown how many customers GOG has, but they clearly have enough to keep existing for almost 20 years while selling games for as low as €2. Even if they are not as big as Steam (let's say: 20 million), if even 5% of these people are using Linux, it's still another 1 million users.

Slowly but surely, Linux is getting into the realm where the platform can't be ignored anymore on desktops and laptops.

[-]

Patient_Sink@reddit

Sure. There are also people on Linux not gaming at all, and of course the same thing applies to windows and macOS as well. But ops thing is only centered around steam, so it's kinda irrelevant for their point.

[-]

DestroyedLolo@reddit

Ok, you're right : time to switch to NetBSD :)

[-]

HexspaReloaded@reddit

15% final seems achievable

[-]

PossuTryffeli@reddit

I am going to mark my words here and say that Linux will get to 10% before 2028.

[-]

Journeyj012@reddit

I'm going to dip my toes in here, and say that the only way we get that close is if we recreate Excel, but only if it's perfectly compatible with the Microslop version.

[-]

PossuTryffeli@reddit

I would say that Euro Office is in a good shape after the argument with Only Office.

[-]

BashfulMelon@reddit

This is not a serious claim because predicting things is hard, especially those that take place in the future.

However, it tells me what I want to believe therefore it's real and true.

Do desktop environments next. I want to see everything that isn't KDE Plasma in the negatives. Thanks in advance.

[-]