Apple-Intel divorce to be final next year Apple's macOS operating system will drop support for Intel chips next year, marking the end of a twenty-year relationship. At its Worldwide Developer Conference 2025, the iBiz confirmed that its forthcoming macOS 26 Tahoe release this fall will be the last to support Intel chips. "Apple Silicon enables us all to achieve …
Aren't you forgetting all those Raspberry Pi and other ARM boxen running Linux (and other OSes)?
If I were Intel or AMD, I'd be getting quite worried these days. Even Windows is dipping its toes into other CPU architectures (again)…
And ARM might also see RISC-V come along and take some of its share…
Raspberry Pi's run a proprietary OS on the GPU - GNU/Linux is the secondary OS.
ARM SoC's aren't really as useful as real computers - there are many more real computers that either intel or AMD, AMD64 or i686 running GNU/Linux or GNU/Linux-libre.
AMD and Intel aren't worried - as the AMD64 CPUs they sell are still generally the fastest processors available and the low-power models are now reasonably efficient enough for everything but demon rectangles.
Not really. In all the surveys over several decades: Windows on Intel has been overwhelmingly the dominant platform with Apple accounting for circa 10% and everyone else less than 5%.
As for Windows on ARM etc. until there is a catastrophic event, don’t see this as being anything other than an attempt to discourage Non-Microsoft on ARM growing out of the phone market.
Thus as a serious user you have few real choices. A few governments could take China’s lead and major on a different OS and thus encourage standardisation of platform around that OS (ie. Instead of Windows compatibility being the primary consideration, it becomes a secondary consideration).
>In all the surveys over several decades: Windows on Intel has been overwhelmingly the dominant platform with Apple accounting for circa 10% and everyone else less than 5%.
Yes, in measurements of browser useragents for the visitors to a limited subset of websites, or in surveys to a limited subset of people.
It also needs to be considered that any OS can pass any useragent (firefox in anti-fingerprinting mode sometimes uses the windows useragent when running on GNU/Linux for example), which would skew the results too.
But such measurements only confirm that most of the users of such websites run windows and that most of those surveyed run windows - it is fraught with error to try to apply such statistics to OS usage as a whole at a percentage.
The dominant platform on such websites is now Android, followed by windows, followed by iOS; https://gs.statcounter.com/os-market-share/ (although the stats are quite suspicious, with a separate "OS X" and macos count?).
GNU/Linux is being used extremely often now - but it appears that it is not being used to browse such kind of websites.
>Windows compatibility
Those 2 words that don't really belong next to each other.
You are perfectly welcome to get yourself something else.
I know what Apple hardware actually saves me so it's a price I'm willing to pay. And that 2019 Intel laptop with USB-C I had before? That's still in use, and still gets updates. So that's 6 years of use out of a machine and it's still working perfectly. Ditto for an iPhone 6 I sold a year ago: still working, still received security updates. And no, I don't buy an iPhone every year - I have an iPhone 14 which still works fine.
If that situation changes I can switch, I also have two machines running Linux, but Macs make for very usable, stable machines that just work. Expensive in CAPEX, but from an OPEX perspective they more than make up for it in usability, functionality and security, and as long as they keep doing that I will keep buying them. Simple.
"Apple's chip development progress is nothing short of miraculous."
Then they put them in cases with little to no cooling, limiting both performance and the life of the 'miracle' chips. This is Apple, asking for top-dollar for hardware that looks nice, but contains cheapest bidder Chinese back-alley product.
https://youtu.be/SSUbewBAnJs?si=h18-F5AZbGU0kno2
On the one hand, if you bought an Intel Mac two years ago, this sucks.
On the other hand, if you didn’t notice the writing on the wall - written more than five years ago - what the hell were you doing dropping so much money on a computer? It’s like buying a house when you know that it has serious subsidence and is built on an ancient cemetery.
I bought my MBP Intel 2020 in anticipation of Apple having to run through several iterations with the new silicon and OS to get things right. They've done some interesting stuff with the chips and memory and, for many, they produce the ideal machines whether it's for development or "creative" stuff and many people will stick with them a result. My next machine may well be a Mini, with the MBP reserved for travel and Windows VMs. But I haven't been in lockstep with their OSes for years: too many bugs in the new releases and too many API changes requiring new versions of the software.
I've no interest in a Windows machine and this would be for development anyway. This is actually an area where Linux sense, as long as I don't have to deal with any GUIs! I've got several RPis doing various things, so that's also an option. But, as you say, the Mac hardware seems to just run and run, which generally means less time tinkering with it.
I have one Intel machine from 2007 and one from 2011, both running Linux, both doing useful work every day. Like Trigger's broom, neither is exactly as it was when new, then again, that's not even an option with a Mac. Neither would be much good at matrix inversion, but it's seldom I really feel the desire for something faster. If it ain't broke...
... And thanks to OCLP, I plan to use it (as a Mac) until fall 2028, so 10 years.
And beyond fall 2028, thanks to Windows Server 2022 with desktop experience, it will be a windows machine until ~2033
And beyond that Linux Mint T2 will keep it running, and out of the Landfills for many more years.
Yes, at some point it will be handed down to someone else, but having options is nice.
In another note, 2028 will also mark the aproximate year when my Synology DS1515+ will become unsupported (after hacking it to accept DSM 7.2) So, it will be the year of reckoning, buying new desktop, Laptop and NAS.
Thing is, if Apple and Synology do not change their ways, I may end up buying some other brands.
PS: To sysnology, get moving FAST with 3rd party disks and NVMe in the compatibility list, to apple, make at least one desktop and one laptop that can get upgradeable memory and upgradeable NVMe drives in industry standard formats, should not be that hard if you do the MxPro with integrated memory, while the Maxs (and ultras) use CAMM2/SO-DIMMs
But from here on to Fall 2028, may things can happen, so I'll think about it in 28H1
It'd help if Apple were forced to release the full specs of the T 'security + lots of other things' used on the Intel Mac* rather than forcing people to reverse engineer them
You should blame the chip vendors on that one.. (Yes, I know Apple is secretive but Intel et. al. are far from open books - especially with stuff like the secure enclave technologies..)
Being able to run bootccamp (and therefore windows on bare metal for things like gaming) and being able to run VMs with Windows exclusive things (like SAP clients or Visio) is a boon for certain types.
Having said that, that was from 2006 to 2018. With the advent of Win10 for ARM in 2018, Windows AMR VMs can be had, but not all the SW is still ARM (Visio probably is, SAP is still not). So, nowadays, less than before.
There were reports at the time of Steve Jobs joking/trolling that Apple would be delighted to run windows on intel macs, but Windows could not do UEFI and was stuck in BIOS land.
Guess I need to dye my grey hair. My last Apple computer purchase (1993 Quadra 610 with Motorola 68040 CPU) could read 3.5in MS DOS disks and would mount them with an icon with "DOS". I could read/write to the disk and move files to PCs as needed.
Yes, you've got me there. Bootcamp was a valuable thing for some users, but I still submit it is not a feature that a majority of users use or one that Apple touted as a selling point.
I'm not a Mac user, so forgive me if I'm wrong... the last Macs that use x86 were released in 2019, is that right?
Bootcamp was a valuable thing for some users
I've got 2 work Macs - an old Intel one (about 7 years old) that won't get updates or be able to run Tahoe soon - I'm planning on seeing if I can buy it from the company as I've got bootcamp on it (also Parallels Pro - I've got a couple of work-build VMs for doing Windows-type stuff).
My other is a recent Aplle Silicon MBP. Lovely machine and, once work do an ARM-Windows build I'll do the same thing with Parallels in it.
"How much does the *A* in *ARM* pay in IP royalties for its cores?"
Apple isn't the A in ARM. That was "advanced" and before that, it was "Acorn". Apple was there when the corporate version of ARM was created, but they didn't stay with it all the way through. They are a licensee like others, and while they probably got a discount on their license by planning to build so many of them, they got the full-strength license which lets them design their own cores, so the answer to your question is probably a lot plus the cost of doing the design themselves.
"That is the big draw of RISC-V; royalty free designs for all."
You are mistaken. The ISA is royalty-free. The designs are covered by whatever terms the designers put on them. If you want to use a core today, not by designing your own, then you'll be paying royalties to whatever design company you select for that chip design. The difference is that the design company didn't have to pay the RISC-V people to make it.
Imaginary property does not exist; https://www.gnu.org/philosophy/not-ipr.html
You are thinking of soft cores.
RISC-V also offers reference HDL designs under a free license, gratis - so to copy any of such designs would not require royalties.
Typically businesses take the reference RISC-V designs and add their own proprietary extensions and speedups rather than contract such process out.
Given that I was not talking about IP, but what actually happens out here in the real world when you have some hardware with a RISC-V chip in it, your post is irrelevant. If I was talking about IP law, your post would still be irrelevant, because it's an opinion on what they dislike. Here's when that's relevant: when we're having a discussion about what we want the law to be, what we would write if we had that power, or what we're campaigning for. We are not having that discussion.
You wrote about imaginary property royalties.
In the real world imaginary property does not exist - there is trademark law, copyright law, patent law, trade secret law, contract law (none of which are property laws, most are rather monopoly laws) and many others that are all completely different.
The only thing that needs campaigning against is such popular error, as there is no point campaigning against imaginary laws.
As far as I can tell, with hardware fabbed to make a RISC-V chip, most businesses that sell such hardware, appear to almost always use the RISC-V reference designs and don't pay royalties for modified designs.
I wrote about royalties, the kind you would actually have to pay in real life, which is what we were talking about. You introduced the useless "imaginary property" term, which would have just been annoying if I had prompted you by using the actual "intellectual property" term. I didn't use that term. Therefore, your comment was not only annoying, it was also completely unrelated to anything involved. If I didn't say it was property, and I didn't, nor have I yet, then pointing out that it's not property is not a useful response. If these pedantic little points are so important to you and they probably shouldn't be, you could at least look to see if the person said what you want to argue against before you jump right in.
What I said is that you do have to pay the people who made RISC-V cores when you want to use them, and you do, and that's important if you think RISC-V means you get the practical designs without payment because you'll be surprised when you have to pay the core designers for that. Part of the reason you, personally, will be surprised is your mistaken assumption:
"As far as I can tell, with hardware fabbed to make a RISC-V chip, most businesses that sell such hardware, appear to almost always use the RISC-V reference designs and don't pay royalties for modified designs."
Look harder. SiFive, Alibaba (Xuantie), and SpacemiT are all producing their own designs. When you find RISC-V hardware for sale, what chips are they running? There are a lot of variants, but three of the most common sources are...SiFive, SpacemiT, and Alibaba. Not that those are the only places making their own designs. Those designs are not available for free, either the common definition of not having to pay or any license with some or all of the spirit of free software or open source. Nothing stops someone from making ones that do, and there are people who are doing it, but the fastest implementations that get used when people want performance aren't those.
I don't think you're correct. Apple and ARM announced that they had signed another license only a few years ago. According to their own announcement:
1. Apple had to renew a license, meaning they did not have a perpetual, everything for free license before or they could have kept it.
2. The license they signed ends in 2040, suggesting that they will need to relicense then if they haven't switched to something else, so they don't have one now.
3. It doesn't say what was paid or how regularly, meaning they may have a large recurring payment requirement.
In addition, ARM makes new things. Does your theory account for what happened when ARM released ARMV9 and Apple started using it? Qualcomm doesn't seem to have that either, or some of the cases regarding their use of Nuvia's licenses wouldn't have gone the way they did.
The long standing, admittedly insider, joke about Apple being the A in ARM stems from them buying a 40+% interest in Acorn to finance and secure the supply chain for the Newton and form ARM Tech Ltd.
That 1990 investment saved what would become ARM from the scrap yard, where it was heading fast.
This yielded, among other things, the long term royalty free IP deal, which also permitted the A-series SOCs to vastly outpower their competitors; and the ability of Jobs to perform a seemingly miraculous turn around of Apple -- based on selling of the post-1998 IPO stake in ARM for some 38x what it paid -- and fuelling it ascendance.
Why would Apple want to stick with x86 and simply trade horses? Their CPUs are superior to both Intel *and* AMD in single thread performance AND performance per watt. The only reason they don't beat them in multithread as well is because Apple doesn't make CPUs with as many cores as they do.
> Apple generally makes security updates for macOS available for three years after release. So Intel hardware under macOS 26 Tahoe should linger at least until late 2028.
This time around, there's no "generally" or "should" required, as Apple has taken the rare step of explicitly stating it:
> macOS Tahoe will be the last release for Intel-based Mac computers. Those systems will continue to receive security updates for 3 years.
While I'm sure this was an important part of the decision to move to ARM, there were probably several other reasons:
1. iPhone uses ARM, and Apple already had a license to make their own ARM cores.
2. They can make ARM processors with exactly the features they want, where they would have to use off-the-shelf Intel processors.
3. Having ARM on laptop and desktop computers makes development of software for iPhones easier.
I was working for a hardware/software company that specialized in digital signal processing. A lot of signal processing. They liked to use PowerPC chippery, because their floating-point units were superior to Intel's.
In the mid-noughts, they started drooling about new processors being designed especially for this niche by a little-known start-up called P.A. Semi. They had early prototypes and good rapport with the chip designers.
Then Apple came in, and bought P.A. Semi, not for its products, but for their expertise. The engineering team started working on what is now Apple's own designs, never releasing their own products.
P.A. Semi were set up by engineers who created the DEC Alpha and the StrongARM who left when Intel aquired the chip business. These are the chaps that really knew how to get the best out of the ARM architecture, so it was no surprise that Apple took them in-house to design their new silicon.
It's very satisfying seeing Apple back using the same architecture they partnered with Acorn to bring to their earliest hand held devices, and creating some of the best performing chips in class.
I'm getting a bit fed up with reading that Apple invented the smartphone, as strongly implied in this article:
"Performance per watt had become an issue around the turn of the millennium in datacenters and in high-performance computing. And it became a broader concern with the arrival of the iPhone in 2007 and the Android devices that followed. "
The first smartphone I ever saw was a Nokia communicator. The 9210 was ARM-based in 2001. By 2003 Palm's Treo had followed suit. The iPhone came out in 2007. Android followed in 2008, running on hardware by HTC, which had been selling ARM-based Windows Mobile devices since 2004. These all used ARM for its low power consumption, as did Apple's own Newton the previous decade for the same reason.
I get that the iPhone ushered in the mass-market adoption of smartphone, subsequently capitalised on by Google. But that was much more to do with Steve Jobs' hardball negotiations with carriers over the price of cellular data than anything else.
-A.
In defense of the article, many of the phones that came before the IOS and Android wave needed low power but didn't need or, in many cases, have good performance. Larger mobile OSes, running applications in an environment that allowed them to do more things, meant that there was a substantial increase in the need for CPU power without raising the power consumption very much. Many of the operating systems that came before had limited multitasking, either because they hadn't bothered to write it or because they wanted to have good performance out of relatively restricted hardware. Windows Mobile was to some extent an exception there, but IOS and Android did see an acceleration in resource usage. I'm not entirely sure that's what the article intended to say, but it is the most accurate way I can interpret their statement.
Well yes, if you like. I would say that mobile data (GPRS) was so expensive back then, and so poor in performance and coverage that just getting the thing to do anything at all useful was pretty magical, no multitasking required. I remember having to get my Treo 270 (Dragonball, not ARM) to connect to dial-up servers over 2G and being thankful for any kind of online connectivity at all, at the per-minute cost of a phone call, because 2.75G wasn't available where I happened to be.
Proper multitasking,well, that's good, right? And yet I can't think of a compelling use for it on mobile. When I'm using an app in the foreground I really don't care about what's going on in the background. In fact only yesterday I grabbed my remaining Android device by the throat and turned off every single notification after it woke me in the small hours.
-A.
I appreciate having some things running simultaneously. I can have a navigation app with directions, send an email, and go back to the navigation app to find that it still knows where I am and where I'm going. If I had the voice instructions enabled, it would have read them out while I had the email program open. Without multitasking, it would stop working until I reloaded it, most likely parsing the map data from disk again and, in my experience, placing me back on the main screen. I can use multiple messaging apps and receive messages on any of them, meaning I can use encrypted chat with people who want while keeping an unencrypted version for everyone else. Apps that are downloading something don't have to stay in the foreground, and ones that are set to do something on a schedule don't need me to awaken them to do it*.
My response was about more than multitasking, though. I also agree that data speed was an important factor, because it made many applications viable which were never written for earlier operating systems. So did many other increases in hardware enable software which people wanted but could only really run on the more heavily resourced smartphones that came after the iPhone. Without the iPhone, that would almost certainly have happened anyway with other companies building the thing. Apple can take no credit for causing that, but if they want to, I think they can fairly take credit for being one of the most influential manufacturers around when technological development made it happen, hence why the iPhone's design** has had a lot more effect on modern phone design than Nokia's or HTC's 2007-era smartphones.
* Yes, I effectively run cron jobs from my phone. Weirdly enough, there do seem to be a few gaps in both IOS and Android's ability for apps to run in the background all the time. All the important things I've done with it have worked flawlessly, but a few unimportant things seem to have hiccups occasionally. For example, I listen to podcasts and I want my podcast app to refresh and download them overnight, which it usually does, but occasionally, it hasn't for some reason and I don't know why. So it's not perfect.
** I refer to UI design, OS design, and hardware design. It's remarkable how similar modern Android phones are to the first iPhone.
I think the iPhone gave the smartphone industry the kick up the arse that it needed.
As someone who had for years been happily(ish) using a series of Windows CE PDAs and smartphones (from the likes of Dell, HTC, Motorola etc.), I was blown away when I first saw the iPhone. At the time I was using a Motorola Q, which I was really happy with and considered it pretty much the best smartphone I'd ever had, but after playing with a friends iPhone I felt like I was stuck in the dark ages. The iPhone was the full web and email on a phone, not the broken, only half working stuff I was used to on Windows CE. And navigating with a finger was so much more fluid than a tiny stylus or keyboard. I still didn't buy an iPhone for a few years, but it certainly opened my eyes to what was possible.
Just take a look at the early pre-release builds of Android that were being shown off at the time. It was far more akin to Windows CE than it was to iOS and was built around devices with keyboards and a D-pad. Android had to have a rethink after the iPhone was released, which was a good thing for everyone involved. Let's face it, Android surpassed iOS in functionality pretty quickly (copy and paste anyone?).
I'm sure the industry would have gotten to where we are sooner or later, but I'd give Apple some credit for massively pushing things forward at a time when most were using Windows CE - which had been stuck in its ways for years - and pre-release Android looked like it would be more of the same.
I'm getting a bit fed up with reading that Apple invented the smartphone
I totally agree. What Apple did was (a) simply make it far more user friendly and (b) ended the control carriers had over allowing phones on their network or not, but smartphones were there way before Apple came to the market.
Technically, even the Motorola V3i was a smartphone. Stupidly shiny keyboard (the matt versions that came out later were far more usable) and terrible to ENTER information, but OK-ish at the time for reading small emails and IMHO just about the best form factor ever for a phone.
As for keyboards, it is worth mentioning the Sony Ericsson P1 who applied a very clever trick to get away with a 20 key keyboard: the keys could be pressed on the left or the right side and so offered the equivalent of 2 keys. Still a bit cumbersome, but even usable for someone with my thick fingers :).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
