I’ve long lost the ability to keep track of whatever’s happening in the Amiga community, and personally I tend to just focus on tracking MorphOS and AROS as best I can. The remnants of the real AmigaOS, and especially who owns, maintains, and develops which version, are mired in legal battles and ownership limbo, and since I can think of about a trillion things I’d rather do than keep track of the interpersonal drama by reading various Amiga forums, I honestly didn’t even realise there’s been a development in the Hyperion Entertainment situation.
Hyperion Entertainment is the Belgian company who has been developing both AmigaOS 4 and 3.1/3.2 for a while now, but the company’s largest shareholder, Ben Hermans BV, went bankrupt, causing its shares to be annulled as prescribed under Belgian law. This happened well over a decade ago, but only earlier this year, in January, was the situation resolved for Hyperion: a new director, Timothy De Groote, was appointed by the remaining shareholders, who also instructed Hyperion to continue development of Amiga OS.
In addition, a few days ago, Hyperion released Update 3 for AmigaOS 3.2, adding a bunch of fixes and improvements to AmigaOS 3.2.2. It brings various updates to ReAction classes, a new custom menu for TextEditor users can customise with macros, a new KickStart 3.2.3 ROM, and many more smaller updates and fixes. The update is free for existing users. AmigaOS 3.2 is available for classic Amigas.
ACube Systems, which makes the SAM460 line of motherboards used to install AmigaOS, just announced a firmware update. So there is still limited life with the Amiga:
https://www.acube-systems.biz/index.php?page=news&id=163 April 8, 2025
They make the following claim:
“For the first time in Amiga history, it is now possible to boot from an NVMe SSD.”
I just use my Amiga and try not to think too much about all the other stuff… That being said I did put 3.2 roms in and I will get some 3.2.3 roms at some point. I prefer to do that than use patches that have to load at boot time.
I love my 3.9 setup, i am not ready to change just yet.
I like 3.9 too, However I found it slightly unstable for me, It was before I recapped though so maybe it would be fine now… But 3.2 is fine and seems good enough for my use.
I’d actually buy AmigaOS 4 if it was available for x86 architecture. PPC is old and loud (as any G5 PowerMac proves) and I don’t really have space to store another computer in my tiny apartment.
AmigaOS 4 will never be available for the x86 architecture because it has lots of handwritten PPC assembly in it. It’s like hoping that MS-DOS (or FreeDOS) gets ported from x86 to PPC (or ARM or whatever). It ain’t gonna happen. And if it did happen, we’d be talking about a different OS altogether.
That said, AROS is an AmigaOS-like operating system for x86 if you are interested.
Microsoft Windows has a lot of x86 assumptions (and legacy baggage) but NT was available across a wide range of architectures. Similarly, MacOS X has withstood 3 architecture migrations (PPC/x86/ARM). It’s not impossible, but it might be impractical. It’s of course worth remembering that the Amiga, like the Macintosh, was born on 68k and migrated to PPC.
Though i get your point. It might be more accurate to call an x86 AmigaOS port “AmigaOS 5” to distinguish it from AmigaOS3 and AmigaOS4. but it’s rather unfair to call such a port “a different OS altogether”. I don’t think you’d call Windows 11 on ARM a “different OS” to Windows 11 on x86, or MacOS on ARM different to MacOS on x86.
When it comes to Windows NT, definitely not, because it wasn’t even initially written for x86, it was written for the Intel i860 (codenamed “N10”) and then ported to x86, precisely to avoid x86 assumptions from being introduced into the codebase. It’s right there in Wikipedia:
https://en.wikipedia.org/wiki/Windows_NT
Fun fact: The codename N10 (“Nten”) would eventually form the product name for the new OS (Windows NT). Of course, Windows 9x does have lots of x86 assumptions and never left x86.
So, saying “Microsoft Windows has a lot of x86 assumptions” is highly ambiguous (and dare I say a bit disingenuous) because it conflates Windows 9x and Windows NT
Yes, because MacOS X was designed to be portable from day 1, with Rhapsody being available for both PowerPC and x86 right from the beginning (Rhapsody DR 1 / Rhapsody 5.0). Again, it’s right there in Wikipedia:
https://en.wikipedia.org/wiki/Rhapsody_(operating_system)
Rhapsody 5.3 dropped x86 support and was released commercially as MacOS X Server 1.0, which later was developed into MacOS X Public Beta and then MacOS X 10.0, but the OS itself was portable from the start.
It would require rewriting extensive parts of the OS and removing any baked-in PPC assumptions, to the point you would essentially have a new OS, even if it’s (mostly) API compatible. In the same way Windows NT is a different OS from Windows 9x even if it’s (mostly) API compatible with Windows 9x apps.
And I say mostly API compatible because certain Windows 9x assumptions such as assuming a 512MB memory limit or the absence of SMP or that weird thunks to MS-DOS are possible simply cannot be reconciled by API compatibility (and some Windows 9x software chokes on it when running on Windows NT). A “new Amiga OS” will have similar issues with its PPC platform assumptions.
Although on second thought, instead of “PPC platform assumptions” I should have said “old AmigaOS assumptions”.
And let’s not forget the main issue here, which is that Hyperion doesn’t have the resources to do such massive rewrite. Microsoft did it because the Windows API was huge, so there was enough demand for a ground-up new OS that is (mostly) backwards-compatible with the Windows API.
@kurkosdr
“AmigaOS 4 will never be available for the x86 architecture because it has lots of handwritten PPC assembly in it.”
That is not even close to being true. I’ve seen you make this claim on here before. Why?
My brother is one of the OS4 developers, here’s what he just wrote when i showed him what you wrote.
“ Not remotely. AmigaOS4 is almost 100% written in C. The only assembly language are in predictably low level operations that have to care about the CPU that any OS has. Stuff like knowing the stack layout, exception traps, special features, MMU etc.
One or two functions optimised for performance, e.g. memory copy and such
The whole reason it’s written in C is because that’s what it took to get off 68K.
Who is this muppet?”
Then I guess they are tied to the PPC in other ways, such as making too many assumptions that tie them to certain PPC platforms? I mean the FreeDOS kernel is also written mostly in C, but good luck porting it to anything other than x86.
Nope. It’s stuck on PPC due to contractual things and lack of money not technical reasons.
Except AmigaOS was already ported from m68k to PPC, there’s no reason a port to x86, ARM or RISCV couldn’t be done.
The decision to stick to PPC was more political and contractual issues than a technical one.
In terms of compatibility issues, the transition from m68k to PPC introduced compatibility issues too, so transitioning to another architecture wouldn’t be any worse.
AROS have you covered, SMP wa sloved long ago and so was the memory limits and restrictions.