Not sure what your point is here. How does this demonstrate Apple's influence on the industry? These are all Apple machines, some more successful than others (two only lasted a single generation).
The point I am getting at is that Apple eventually got SFF & PCIe slot-free commercially successful.
It's also worth noting that despite the iMac's 'success', AIOs have never been a thing in the PC market, because other options exist.
All-in-one (AIO) PCs are one of the major product types of desktops and their shipments will amount to 11.06 million units, accounting for 12.8% of the global desktop volumes in 2022.
The PC industry is having a rough 2022 with factors such as global inflation, the Ukraine-Russia war, COVID-19 lockdowns in China, and increasing interest rates by mature markets all significantly undermining the global economy, resulting in PC demand shrinkage, but the situation is likely to...
www.digitimes.com
By comparison PC workstation that typically use Xeon, EPYC & Threadripper chips shipped
<7.7 million in the same year.
OK, but you're the one that brought up Grace Hopper. I assumed you were referencing its overall architecture, as shown in your diagram, rather than suggesting Apple makes a $100K Mac Studio.
The point I was making at is the concept is sound and people critical of it are not forward thinking unless their darling brand does it themselves.
OK, but in that story x86 was the underdog, offering just enough performance at a much lower price, displacing the big-iron heavy hitters. ARM (or Risc-V) overall may be on a similar trajectory, but Apple Silicon certainly isn't cheaper than x86.
Macs weren't cheap using x86 either.
At the same/similar/lower price point as Intel Macs, Apple was able to improve battery life, power consumption, size, weight, thermals, form factor, operational noise and other quality of life improvements lacking with x86 counterparts.
I think you'll find that x86 obsoleted 68K and PowerPC, not Apple. In both cases, Apple had no option but to transition (the second time to x86), or become obsolete themselves. Same with OS 9, which was absolutely creaking by the late 90's, after Apple failed several times to come up with a replacement. If they hadn't bought Next, Windows 2000 / XP would have buried Apple. OS 9 didn't even support pre-emptive multitasking.
I am pointing out that Apple is not loyal or emotional to any specific tech. When they see that those now defunct tech were starting to ascend to the top of their S-Curve they start looking for other alternatives that are at the bottom of theirs and will outpace the incumbent.
Apple's RISC-V job position may be a sign of what is to come 1-2 decades from now.
“There’s an old Wayne Gretzky quote that I love. ‘
I skate to where the puck is going to be, not where it has been.’ And we’ve always tried to do that at Apple. Since the very very beginning. And we always will.” - Steve Jobs
If high-end users want raw performance, why doesn't Apple just give them raw performance? Why are they 'evaluating the future' and providing efficiency and specialisation instead? High end users would happily use less power, but it's not their prime consideration.
There are a couple of ways to look at this. Apple failed with the M2 Extreme for fab reasons and they will make another attempt in Q1 2025. The other would be the Mac raw performance users aren't numerous enough to take the effort. No one can satisfy all users all the time. If Apple can address ~99.99% of all typical Mac use cases within their sweet spot then it is good enough. Try again next gen that has been observed to be every 19.5 months.
This is unlike x86 where in they have <7.7 million PC workstations shipped last year. This is why it is often pointed out that maybe it is time to switch?
Bear in mind that Intel's best desktop / laptop chips outperform ASi whilst being on 10nm, whereas Apple enjoys the advantage of 5nm production. If / when Intel sort out their production, the gap could close pretty rapidly. A 5nm 13900K would be impressive.
That isn't the direction of the PC industry as a whole. Today and into the foreseeable future it is efficiency and specialization.
Why is Nvidia producing ARM SoCs?
No one's demanding macOS on x86 or RTX. Ultimately, it's just a cost-benefit of computing power x OS quality x price. Apple have a great OS, and in general make nice machines. Their laptops in particular are compelling. The issue with their desktops are a) a lack of user upgradeability / serviceability, b) a lack of high end GPUs, c) prices that start reasonably (apart from the Mac Pro), but skyrocket with spec upgrades.
Look at any thread concerning the M2 Ultra you will see users demanding i9 and RTX for their ability to run synthetic benchmarks far better than any Ultra chip could hope for.
Whole PC market has workstations being less than 2.62%. Macs shipped 28.6 million units worldwide in 2022. That less than 2.62% translates to ~75,000/year pro desktop Macs. ~60,000/year are Mac Studio & ~15,000/year are Mac Pro.
20% of Mac Pro users who demand swappable parts and separate CPU & dGPU are equivalent to ~3,000 units annually.
This is the sad reality of capitalism at work that keeps people paid for the goods and services they provide as product refreshes and product support.
How does the economies of scale workout for ~0.01% of Macs sold annually? Swappable parts does not translate to any direct sales revenue for Apple. Mac Pro users via OCLP can extend the useful life of their Mac Pro from the industry observed standard of 4-6 years to 14-16 years or even longer.
Someone pointed out that
large install sites like those at NASA have scientists enjoying the services of dozens of 2012 Mac Pros until today. That's decade+ by now.
The market is growing smaller per year because cheaper alternatives that are "good enough" have typical tower users moving to these other devices. With the niche approaching that of audiophiles at best and mainframes at worst... what to do?
As early as 2020 Apple abandoned x86 for that reason. This has embolden
Qualcomm, Microsoft and other ARM SoC brands to try their luck on Windows 11/12 on ARM another go. Why include Microsoft because they missed the bus with the iPhone, iPad and Macbooks that they're tryng to make up for with Microsoft Surface devices. x86 is not very good in those form factors.
When
ARM laptops become as wildly successful as Macbooks what will happen to legacy x86 as a whole? ~80% of all PCs both Mac chips & x86 are laptops. Where will R&D money for Xeons, EPYC and Threadripper come from? ~90% of Mac chip R&D is from the nearing 300 million annual iPhone/iPad sales. That's the equivalent of all x86 shipments from both AMD & Intel excluding Macs. Android ARM R&D is from ~1 billion smartphones shipped annually.
A repeat of this conversation will likely occur before the 2030s but with Windows as those Android ARM SoC are also all about efficiency and specialization.
