Decline often happens slowly, gradually and then suddenly. Could anybody imagine Intel where it is now ? This could happen to Microsoft and is probably already happening as we speak.
developers don't control what platforms an enterprise would use. Vendors don't dictate the platform either - vendors sell to a customer, and so it makes sense that the customer dictates the platform.
when migration to different platform happens, it's because there's something disruptive that enterprises need to move to, or a new class of enterprise without existing/legacy baggage sees competitive advantage in moving. This happened to IBM when their mainframes no longer offered competitive advantage over the newly minted PC platforms.
If/when windows become lackluster in terms of a required feature, or did not keep up with a needed feature that an alternative platform provides, then the switch will happen fast. What that feature might be i dont know - if i knew, i'd be making it.
They might not control it directly, but they absolutely influence it. Linux was on the losing end of this for many years, as common end-user enterprise software was native and only available for Windows (or in the case of Microsoft Office, nominally available for Mac OS but with fewer features and lots more bugs). That was Microsoft's moat and it started leaking when web applications became ubiquitous. That leak later accelerated when those web applications had to work on mobile operating systems (namely iOS and Android) that Microsoft did not own and could not control.
> Vendors don't dictate the platform either - vendors sell to a customer, > and so it makes sense that the customer dictates the platform.
There are plenty of counterexamples here. I used to have two legacy SGI machines in my cubicle at work precisely because a vendor dictated the platform to that Very Big Enterprise company many years earlier.
Similarly, many people buy Macs solely to run Logic Pro or Final Cut Pro, because the vendor (Apple) dictated the platform by discontinuing the Windows versions. Apple doesn't have the market share Microsoft has, but unlike Microsoft, they can maintain strong control because of their breadth (OS and hardware for desktop, tablet, and phone, plus high-end creative software) and because a lot of their customers are all-in on Apple's ecosystem.
This comment is multiple years out of date at this point…Lunar lake was considered an exceptionally successful Mobile processor platform that was extremely competitive with Apple M series chips, Panther lake is supposedly improving on that by an even greater degree, and Nova lake is supposedly an extremely promising upcoming desktop platform, and their Arc graphics are excellent for the price point. This is like writing off MacBooks during their thin and hot touchbar/butterfly keyboard era and then assuming they’re that way still in 2026 when they’re now extremely well received laptops lol
Apple barely comes into the picture, with less than 10% of the worldwide PC market, and was not a factor in Intel's decline.
governments, institutions, and large enterprises (like, thousands of people) do not have the power to do anything "suddenly". they have contracts, and cash flow concerns. you cannot suddenly replaces tens to hundreds of thousands of machines.
20-50 years down the road? maybe! they (microsoft) surely arent doing themselves many favors. but they are certainly not in "significant danger" today.
It won't be sudden, until it is
It happened with all the vpn+shared drives buried to just use SPO.
different experience,I guess.
Did your employees got trained? or just sent the link to 3 'online trainings'?
Managers manage to switch to Mac seamlessly. I am sure the rest will follow with cheaper macs now available. And now, with 'office on the web', you can use basic office everywhere. (even on Debian)
office is a tiny, almost negligible, piece of the puzzle. quickbooks, solidworks and other cad software, bespoke software, security software, user management, permissions management (replacing active directory), contractual obligations, the millions of dollars required in implementation, the millions more dollars required for increased user support, and so on...
but, again, just to reiterate: i am disputing that windows is in "significant danger" today.
this is also ignoring all of the critical software that is windows-only (e.g. quickbooks, solidworks, bespoke programs in banks and government).
point is: microsoft is not in "significant danger" today.
user management, file management, security, windows-specific software, auditing requirements, required capital investment, lack of competent linux sysadmins compared to windows sysadmins, and so on.
Wasn't this a microsoft marketing campaign like 20 years ago? Are we still on it?
i would love to see your numbers for this. what does "increasing percentage" mean? 1% -> 2%? 10% -> 20%?
i teach at a college level, in tech, and would estimate ~5% of incoming students have any experience with something other than windows on a pc, at best. outside of tech, i would estimate ~2%.
Rest of the world, depends pretty much even if an half work second hand PC is an option.
And Canadian banks aren't known for moving fast. They are pretty conservative (at least the big chartered banks are).
It means that, today, a lot of enterprises begin pondering the question, and then about a year from now, they start seriously studying and prototyping it, and then "suddenly" in 2029 Microsoft starts seeing a deluge of defections. It means a whole bunch of peopling finishing the conversion all at once, relatively speaking, even if that "all at once" is 3-4 years away.
To put it another way, the thresholds where people get annoyed enough to quit are highly correlated to each other. If individuals on HN are posting "I don't want to switch, I've been working this way for decades now, but Windows has crossed the line for me, I've switched to Linux, and it was easier than I thought it would be", then corporations and governments are having very similar deliberations internally.
This is probably a more accurate model for how "influencers" seem to work than the idea that some crazy guy in your organization falls in love with Product X and evangelizes it internally. I'm sure that happens and is a real force, but this correlation-of-experience effect is probably bigger on the whole. If Product X was good enough to make an evangelist internally, or more germane to this conversation, to make some a mortal enemy of it internally, it's usually because it was a good enough or bad enough product to be able to do that in the first place, and eventually everyone will figure it out in exactly the same way, just later.
20 years is way too large a minimum estimate. If Microsoft responds correctly that might be good, but if they just decide to rest on their laurels and extract whatever value they can out of Windows while they can, Windows would never last 20 years of that. Even the slowest organizations can move faster than that. After all, to cut Microsoft's revenues off at the knees, they don't need to remove every last Windows 2000 server in their backoffice they can't upgrade, they need to cut out just the majority of desktop licenses.
i disagree. unless intuit is also rewriting quickbooks, dassault systèmes is rewriting solidworks, every bank is rewriting their custom windows-only software, every government branch is rewriting their custom windows-only software, etc. and every company is willing to retrain 95% of their employees on a new operating system, have increased support requirements for a few years at least, etc.
not even touching the capital required for such a transition that in many cases has questionable benefits (from a business perspective).
time will tell! i have first-hand experience with how fast banks move, so i will stick by my 20 year minimum. happy to see otherwise, though.
in any case. what i replied to was a claim that windows is in "significant danger" today. it is not.
The market you're describing is real, and very significant—but I don't think it's even a majority of Windows users. If so, it's a small one.
And imagine what even 30-40% of all Windows sales disappearing over the course of 2-3 years would do to Microsoft. To Windows as a platform.
Then imagine what would happen if it was 50-70%.
The former, I would describe as "a disaster".
The latter, I would describe as "apocalyptic". (Y'know. For Microsoft as a company. Not in general.)
Product teams deciding it's easier to ship on + customers having enough linux familiarity (from their other projects).
And the current crop of Microsoft people on the Windows team don't seem to understand building a platform in the way 90/00s Windows teams did.
It's clear MS moved a lot of their smartest people over to work on Azure products.
Up front they won't need to do a full rewrite. They'll only need to make it work well enough under Wine.
At a source level, tools like Avalonia's xpf make porting WPF apps to other platforms easier:
They already have. You can't buy QuickBooks for desktop anymore unless you want Enterprise, the expensive $4k+/year subscription. They dumped the Pro/Pro Plus and moved all those users to QuickBooks online.
And now they've launched Intuit Enterprise Suite in an effort to move the QBE customers into Online. The writing is on the wall there, desktop is going away.
It's also happening in more specialized areas too. I work in waste management/recycling, and this industry was traditionally windows heavy with thick clients on desktops. Even the truck scale software is moving to web interfaces, as are the dispatching and asset management.
OS increasingly doesn't matter for most knowledge work.
Yeah, there are going to be industries that will probably never move, certainly not within a 20 year timeline, but there are a ton that are moving or have moved entirely to SaaS and web apps.
Autodesk Fusion runs in browser, and even SolidWorks has an online version. You were saying?...
> every bank is rewriting their custom windows-only software
This has been happening for a while, actually. Typically they rewrite software as webapps, with a microservice-based backend.
> not even touching the capital required for such a transition that in many cases has questionable benefits (from a business perspective).
Software gets rewritten all the time, often in the "Ship of Theseus" way. The next rewrite will just focus on moving stuff away from MS.
Mostly used for visualisation, rather than real work, especially given the browser limitations.
Not sure about big enterprises, but I already see this happening in the mid-size, non tech company market.
I'm an IT manager and has been a sysadmin/ops for my entire career, and the past ~4 years I've been seeing a pretty consistent shift toward companies my company does business with deploying more and more macs. Windows is still dominant in my industry, but the cracks in the wall are widening. It's gotten to the point that I'm genuinely surprised now when I see Windows when someone screen shares.
Apple silicon is just too good and the generations coming into the workforce now don't have a "default" windows familiarity that we used to have. They're coming in needing to be trained on how to use a PC in general, windows or not, having used nothing but chromebooks and mobile OSes.
Now, Office OTOH is more entrenched than windows. Even the macshops I interact with are all on M365. Macs are managed with Intune, users & SSO with Entra, Defender for EDR, and of course the office apps. And that's why Microsoft probably isn't as afraid as it seems when it comes to Windows. Even without Windows lock-in, there is very real M365 lockin that is far more entrenched than the endpoint OS.
it is great that some are starting, seriously. but windows is not in significant danger as of today.
A couple weekends ago, I made the overdue call to kill my dual booting with Windows 10 and go full Linux. I'd considered finding a copy of the embedded Win10 long term version or paying for the patching. The local account was one of the things holding me back from doing the update. I knew I could muck with things to still have it, but figured it would be yoinked away later. Similar thoughts to updating the old threadripper that no longer qualifies for Win11. The reinforcement came from all the blasted copilot integration -- notepad, paint... just looking like evasiveness was going to be everywhere.
For a long time my 'good' box was Linux, my old box was Windows. So much just works. I still have an M2 with Windows 10 on it, but it is not in any machine right now. Will see if I run out of space and need it before they actually provide something I'd want to even have on my desk.
It's not like apple just deciding that we don't do 32bit applications anymore all of a sudden. However many ads MS will shove in windows, as long as they can run software that they depend on, companies and factories won't care.