Go check MacTracker. A 2006 Core Duo Mac stopped being able to run the latest Mac OS X release with Lion in 2011; you could still get security updates for Snow Leopard, its final release through 2013. That's 5 years with a 2 year buffer for security update support. Compare that with a 2012 model (Ivy Bridge/3rd Generation Intel Core i series processor) Mac, which was able to run the latest version of macOS up until twelve days ago. That's 8 years with a 2 year buffer for security update support.
Go do the research for yourself.
Because the chips are not ready yet. If they were, Apple would've fully replaced the entire Mac mini line and not only the low-end Intel model. Also, the M1 Mac mini would be a no-compromises upgrade (in the way that the M1 2-port 13" MacBook Pro was over its Coffee Lake predecessor) and not one where the maximum RAM is one fourth what it is on the Intel model and where there's half the number of Thunderbolt ports. Clearly the M1 has limitations because it's the best Apple can do right now.
Market caps don't mean anything. Apple doesn't need to throw more money at this; the M1 Macs are a critical success and it doesn't hurt Apple to do things exactly as they currently are doing things. Steve Jobs was a master salesman; under-promise and over-deliver. The Intel lineup was fine for the Mac to move at the exact time that it did and the Core Duo was perfect for all but the Mac Pro and Xserve, which waited for the Core 2-based Xeons to move. Once A15, and all associated advances become closer to releasing, you'll see higher-end Macs make the jump. Odds are decent that the Mac Pro's eventual Apple Silicon SoC will need even more time in the oven.
People who still need x86. Not everything runs smoothly in Rosetta 2. And just because you don't need the things a Mac Pro or a 16" MacBook Pro customer might need doesn't mean that those customers don't need their installed software to be perfectly stable until it makes the jump to Apple Silicon.
There are plenty of people out there who will still be buying Intel Macs throughout this transition for that reason. Furthermore, it's not like an Intel Mac is a bad computer just because Apple Silicon is out. Maybe it's not the most practical decision, especially if everything one would do with their Mac is either Apple Silicon native or runs just as well in Rosetta 2. But that's not all people, especially on the high end.
Apple is going to move the machines over as fast as they can. And that length of time seems like it will be two years.
That doesn't mean that they have chips that can feature-for-feature replace every Intel Mac. They have single-core performance down, but they don't have GPUs that beat out the current discrete graphics on the 16" MacBook Pro, 27" iMac, iMac Pro, or Mac Pro yet. That crap takes time to engineer. They also can't break 16GB of RAM or provide more than 2 Thunderbolt ports. That's critical. That's why this transition will take longer than the Intel one did. And really, those are factors that are specific to this transition (and wouldn't ever stop the previous transition due to not being applicable).
You have different wattages on the M1 chips. That's why the M1 Air and M1 2-port 13" MacBook Pro have different chargers and why the latter operates with greater sustained performance than the former and why the M1 Mac mini, theoretically beats out both of them. Go take a look at Apple's M1 performance per watt graph. The performance scales as the wattage increases.
Not for new Macs. But certainly for Intel Macs as far as driver support in new releases is concerned.
Microsoft isn't a chipmaker that designs its own SoCs and solely for its own products the way Apple does. Furthermore, for Windows 10 for ARM64 to succeed, Microsoft can't be the only company making devices for it. The other OEMs have to buy in as well. Microsoft could decide to make SoCs and license them to the other OEMs. That's a possibility. But for that to happen quickly would entail a ton of things happening.
No, Android CPUs have good performance per watt as well. That's a byproduct of the ARM64 architecture. Android may not use said performance efficiently, but those SoCs are also beasts.
They already exist and they do not perform well. Though a lot of that is the lack of decent x86 binary translation in Windows 10 for ARM64. Otherwise, the performance compared to Apple Silicon is leaving much to be desired. But that's because Apple has heavily customized their implentation of ARM64. Qualcomm and Samsung would need to do something similar and, even then, they won't have going for them what Apple does in that Apple as a chipmaker is only designing for Apple and only for Apple's operating system. That's not something that the others will remedy in four years.
My turn to ask you for a citation here. There's no indication that Apple is ready with anything other than an SoC suitable for an entry level iMac at this point. Otherwise, the 2 Thunderbolt port limit and 16GB of RAM cap on the Mac mini (and continued sale of the higher-end Intel Mac mini models as well as the Intel 4-port 13" MacBook Pro) would lead me to believe that Apple still has a ways to go before they're ready to complete this transition.
They don't spend three years in QA. That's not how that typically works.
A) Based on the strength of this specific citation and my research about how long my 2012 iMac was supported from November 2012 with macOS 10.8 to November 2020 with macOS 10.15 I will concede that it is possible that macOS support for select Intel Macs will lengthen from a conservative 2024 to the year 2028 for desktops like the 2020 iMac. Thank you for giving am uncharacteristically better source of information.
B) The Tim Cook doctrine makes it possible that the chips suitable to replace the remaining Intel Macs have finished with their R&D stage by now. At this point it time they're lined up for QC & stress testing before mass production is to commence in the first weeks of 2021. Manufacturing of billions of items is always done efficiently on a schedule to increase utility of bought capacity at near 100%
iPhone chips do not manifest itself out of the ether. It takes years of R&D before it's ready for sale. As such any other product follows the the same method. The R&D also includes the manufacturing process for the design targets of the chips. Like say the 5nm process employed on the M1. Perfecting 5nm process took years to R&D, test and mass produce the machines to manufacture the chips. That's baked into the total end to end R&D time. If you are implying by R&D as the time Apple takes to manufacture the chips then yes, they can make it quickly once all parts of the manufacturing and supply chain are setup and complete.
Apple planned the transition from Intel to Apple Silicon earlier than last year. I've seen rumours about this transition as early as 2017.
M1 chip is being used with the ~80% of all Macs shipped. From a supply chain, finance, sales and business point of view its release, manufacture and distribute is the least complicated and the most profitable to execute a product launch this way.
Think of the M1 as the "Core i3 of Apple chips". Apple is only selling "Core i3" Macs as it represents ~80% of all Macs shipped. The performance improvement is so astonishing people get confused and think this is the only Apple Silicon chip to ever be made.
M1 is not the "best it can do right now" but the M1 is the collective focus of all of the Mac supply chain.
"Early 2021" will debut the "Core i7 of Apple Chips" (following Apple's product naming conventions it may be called the M1X) that is suitable for four TB4/USB4 port Mac mini and 13", MBP 16", iMac 21.5"/27" make up ~19% of all Macs shipped. Mac event scheduled within Jan-Apr 2021 will introduce it.
"Mid 2021" will debut "Xeon of Apple chips" (following Apple's product naming conventions it may be called the M1Z). This will be used on ~1% of all Macs sold that are the iMac Pro and Mac Pro and as such they're suitable to debut by WWDC 2021 in June.
C) Apple will not meet sales targets if they will not finish the transition in less than 2 years.
People who still need x86 will not be significant number to service further. That is why the $1,099 2018 Intel Mac mini 2020, $1,799 & $1,999 Intel MBP 13" are still being sold. Buy it while stocks are available.
Who would buy a slower & older $1,799 MBP 13" when there's faster & newer $1,299 MBP 13"?
"Practical" is another word for "cheap" or "easy".
If Apple wants to beat estimate then it must be completed by WWDC 2021. If they're running a charity for the feelings of people they'll make it 2 years.
D) Apple has the manpower and R&D money to do everything feature by feature. It's $2 trillion company for heaven's sake!
If Apple started in 2017 would that be enough for your "crap takes time to engineer"?
Business smaller than Apple release products on a schedule. Tim Cook did not wake up 3 days before WWDC 2020 with a revelation of iPhone chips in Macs. That has been on the back burner for at least 3 years.
To scale it down for non-supply chain types think of Apple Silicon chip R&D as going to University. It takes a few years to learn study and get tested each semester until you graduate upon the completion of your thesis. Typical college course takes 4 years before graduating? Like product R&D it takes a similar amount of time. Remember, this is an analogy or better yet a metaphor.
The chips are ready by now. They're just waiting for their turn to get manufactured. These things gets scheduled because any idle time in manufacturing is cost that eats into margins.
Have you considered that Apple has not disclosed future chips as it would kill sales of the current model and give the competition advanced warning on Apple's confidential road map?
The transition will be as short as 210 days or faster.
Tim Cook aint no fool. The bottomline matters. Longer transition will negatively impact Apple's Mac business.
D) M1 chip on the MBA and MBP are virtually identical. The difference in charger has more to do with different displays they have that use varying amounts of power, having a heat sink fan, a larger battery and other components that consumes more power than the MBA. See the reason why MBA has 18 hr battery life than MBA has 20 hr battery life
The point I was making between the difference of chip TDP is that the iMac, MacBook Pro 16", and four TB4/USB4 port MacBook Pro 13"and Mac mini will require a chip with more than 15W TDP. For example iMac 27" used a Core i9 with a 125W TDP. Would the M1 with less than 15W be suitable even if it nearly matches the absolute performance of that iMac?
E) Intel and AMD will not be relevant on macOS as early as 2028. Lengthened it in light of the 8 years possible software update support of the 2020 iMac.
F) I never claimed or implied Microsoft is a chip maker. What I meant was continuing the development of Windows for ARM, create a developer tools to make Windows programs into fat binaries that will work on both x86 and ARM, having a x86 translation program for x86 Windows programs to run on Windows for ARM. Essentially copy pasting Apple's workflow.
G) I never claimed or implied Android chips are not good. I said they do not have the performance per Watt of Apple Chips. I also emphasised that Android chips need not be equal to Apple chips. They only need to be superior in performance per Watt to Intel or even AMD.
H) When has Apple ever disclosed any details about any product that will be launched? Unless of course it's a transition that requires the help of 3rd party developers to complete.
Apple does not publish any sort of product or SoC roadmap. A roadmap would introduce the Osborne Effect or allow rivals a competitive advantage over Apple.
Only offering M1 Macs is a supply chain strategy to reduce cost and time of manufacturing and distribution to its absolute lowest.
The more SKUs a supply chain has to deal with the more complexities and cost that are introduced.
Apple shipped an estimated 18.35 million Macs in 2019. ~80% of that are the Macs Apple refreshed to M1 chips. Assuming units shipped remains relatively the same that would be ~14.7 million M1 chips for 12 months to be used on
- 2 standard Mac mini SKUs
- 2 standard Macbook Pro 13" SKUs
- 2 standard Macbook Air SKUs
Having as few SKUs as possible improves mass production and the assembly line productivity. It also allows for quicker shipping times as these are not BTOs. All those 6 SKUs make up ~80% of all Macs shipped. The ~20% among the M1 Macs are BTOs.
Having the most common parts allows for use of certain parts intended for the product XYZ to be applied on product ABC.
To help explain this further let us go back to the first year of Steve Job's return to Apple. He with the help of Tim Cook cut down on a lot of product lines to make their inventory, manufacturing and supply chain more simpler and efficient.
He reduced it to this.
Desktop
- low end with 3 variations he called "small, medium & large" (i.e. iMac)
- high end with 3 variations he called "small, medium & large" (i.e. PowerMac)
Laptop
- low end with 3 variations he called "small, medium & large" (i.e. iBook)
- high end with 3 variations he called "small, medium & large" (i.e. PowerBook)
Rolling out the M1 first makes Apple's supply chain leaner and faster to help to improve margins and get Macs to people like you faster.
Personal note:
Is English your first language?
Do you work in supply chain, purchasing, manufacturing or part of management?
Last edited: