In vague and uncertain terms a few days ago at the Worldwide Developer Conference, Apple announced that support for Core2Duo Macintoshes is ending, and that these systems will not run the most recent version of the OS, now called macOS "Sierra." – We can presume that it'll still be Mac OS X 10.12, rather than going full Solaris and simply being called macOS 12. In a similar and more sensical (but still kind of unfortunate) move, iOS 10 will not support devices with A5 processors.
After (to some, dramatically) increasing the system requirements for Mac OS X with the release of version 10.8 in 2012, the requirements have stayed the same since. The requirements basically specified fully 64-bit platforms. The earliest Intel Macs, what I sometimes personally refer to as transition-era systems, had a mix of 32- and 64-bit processors, and the ones with 64-bit processors still had 32-bit UEFI firmware.
The new requirements for Mac OS X 10.12 are weird. Apple always states it in terms of models, but the easy way to say it is that Core2 and related CPUs are out. Anything with an Intel "Core i" CPU of some sort will still run the newest version of Mac OS X, and if 10.12 continues the trend set in 10.11, they should be really good at it.
This is a weird line to draw, especially since Apple was selling some budget/education focused Core2Duo systems as recently as 2010 or 2011, and those particular systems do support sixteen gigabytes of RAM and can be maintained with new thermal interface material and storage devices, including solid state drives, which should improve performance and responsiveness dramatically.
Systems like this are still going to be nowhere near as fast as even the Core M MacBook, even last year's model, but they are sufficient for a lot of purposes. They make good basic browsing terminals, Office and iWork should be responsive on this hardware, especially late Core2Duo systems with lots of fast RAM and better nVidia chipsets, especially if a solid state drive is added.
Over the past few years there has sort of been a conversation about what it meant that Apple had been "supporting" some older MacBook Pros so long. Systems like the "Late 2007" MacBook Pro (the one with the failing GeFORCE 8600m GPUs) will have been able to run the newest current release of Mac OS for nine years as of the release of Mac OS Sierra later on in the year. They didn't always do it well, but even without an SSD and physical maintenance, they did run.
Apple gives no insight whatsoever into what their decision-making process is for machines like this. Is it because they tested a bunch of them in a lab and found that they do tasks more slowly? Is it because they want to sell new hardware, and schools hanging onto old Core2Duo MacBooks is preventing them from buying newer machines? Is there some actual technical limitation in this hardware?
I don't think it's a technical limitation. The next obvious one was going to be GPUs that support the new Metal graphics system, or the requirement to have some amount of RAM. RAM usage improved a lot in Mac OS X 10.11, but neither the support for better GPUs (which exist in Ivy Bridge machines from 2012, but not Sandy, Arrandale, or Penryn machines) or RAM support (given that MacBook Airs with 4GB of RAM were on sale until just a few months ago) make sense as the need, especially since several of the latest Core2Duo Macs (the 2009 Mini and the 2009-and-newer MacBook, for example) support sixteen gigabytes of RAM.
I don't know what these particular Macs are like. I've never sat down and used one, so my only points of reference right now are what 10.6 through 10.11 were like on a Late 2007 MacBook Pro and what 10.6 through 10.11 have been like on a 2011 Mac Mini (dual/i5/IGP) and a 2011 MacBook Pro (quad/i7/dGPU). On the Windows PC side of things, Core2Duo hardware is still very usable. I've got a ThinkPad T400 that still browses the web fine, and is my main productivity computer when I'm at home. Relatedly, I still run Office and do extremely light browsing on Windows RT devices, which have 2GB of memory, very slow (although still solid state) storage, and very slow ARM processors.
In general, I don't think Apple has a hard time convincing its real "core" customers, people who profess their love of Apple and who replace their systems every few years, to go buy a new Mac. The people who are hard to convince to buy a new system, or even upgrade to a better used system, are the ones who buy used computers or who bought a Mac on a lark and who don't see the value in needing to replace systems so often. This thought process, I believe, goes along with some others. I don't' agree with most of it, but I do understand that not everybody who wants a mac can afford a new laptop every three years, or even every seven or eight years, or even a newer used system every three to five years.
Mac laptops since about 2008 have become pretty physically durable, they clean up well and in general they age pretty well. Core2Duo processors are, I'd argue, not exactly over the hill yet, and most of those Macs support upgrades.
On the iOS side of things, I think it's reasonable to discontinue support in the OS for "A5" devices. The real problem is that Apple has been selling the iPad 2 and the original iPad mini until pretty recently. The iPad mini was on sale until late 2015, which means it's possible to have bought a new iPad less than nine months ago which as of later this year will not receive any more new versions of the OS.
This creates some interesting situations. When the original iPad, with its A4 CPU was abandoned a few years ago, the software scene immediately passed it up. A whole lot of new software came out almost immediately, none of which works on the original iPad. I fear that with iOS 10, the same thing will happen to the iPads 2 and mini.
On the computer side of things, it's a little more complicated. Mac OS X 10.11 will continue to run modern web browsers for several years to come. Long after Apple stops releasing security patches for the system, Microsoft, Adobe, Google, and Mozilla will continue publishing their major applications for the platform. In the 1990s, this wasn't necessarily a problem, because the Internet environment was so different from today.
Today, Apple literally refuses to patch Mac OS X 10.10, and they've done this for the past several years. Often, the upcoming version of an OS will be patched for a critical privilege escalation vulnerability while it's in the preview cycle, but the current version won't. There aren't currently any network vulnerabilities on Mac OS X, as far as I know.
A lot of people will just continue to use these older systems. I don't blame them, I'm using a similarly old Windows PC, and there's already a large contingent of people who are using older Mac OS X computers. Part of the problem I have with that is that we just don't know what'll be discovered or what other ways to compromise systems. As an example of this, Mac OS X computers were vulnerable to DNS and NTP amplification attacks (this has not been patched in older versions of the OS, but the newest versions of the OS should be safe), but they weren't particularly vulnerable to Shellshock, due to the way Apple structured the OS. That said, Shellshock is an important example, because it became a huge vulnerability which allowed remote code execution and privilege escalation on systems via DHCP, and this is based on code that hadn't changed in 23 years when it finally got discovered, exploited, and fixed, in short order.
In general, m ost computers new enough to have just been abandoned by Mac OS area also new enough to do things like send thousands, probably hundreds of thousands of spam emails, or run cryoptographic processes (such as bitcoin mining or a tor endpoint) on behalf of a remote attacker without the end user noticing.
It hasn't been a problem so far, but I fear that at some point it will be, and I don't think that enough people are shedding these types of systems (or simply keeping them offline) for herd immunity to take effect. We definitely need to get rid of the notion that "UNIX == safe".
It will be interesting to see what happens. Generally, prices on newly unsupported Apple hardware start to fall, but this move is going to make systems like the 2010-2012 Mac Pro and the 2011/2012/2013 MacBook Pro systems that much more valuable, as people start to try to buy more of that hardware.
Sometimes I wonder what the actual market for people who want a commercial UNIX desktop but don't want to be in Apple's hardware cycle looks like. Among the people who express this desire, I only know of a few people who have done anything about it. Some of them just run Windows or Linux/BSD systems. One of them bought Solaris licenses. I know a few people who have built Hackintoshes, but that's going to be variably successful on laptop hardware.
None of what Apple's doing by periodically removing support in the OS for older hardware is particularly new or unique to them. Sun killed off support for old machines on a regular basis, Apple did it a lot with Classic Mac OS in the '80s and '90s, and Microsoft has been slowly increasing the requirements for Windows.
It's not straightforwardly clear what Apple's motivation is here. It would be easy to say that it's to sell newer hardware. That's probably part of it. I suspect that there are still internal constraints on Apple's engineering and testing capacity. Even though they've got this giant pile of money, they may not have been able to or wanted to hire enough of the right type of employees to properly test the new OS on systems, some of which will be nine years old.
I hope we'll hear more about exactly what inspired the move, but I doubt it'll happen, especially with Penryn appearing to be a pretty random line, drawn in the sand at about 20009 with no discernable reason. In general, Arrandale isn't a whole lot faster and Arrandale systems didn't have much better graphics hardware.
I suspect we will find out eventually. It either will or will not make any sense at all.