Haha I love these wild guesses. Stabs in the dark.
Could provide chips for iForce One in 2052
Chips for iPlanet in 2342
Could provide chips for iForce One in 2052
Chips for iPlanet in 2342
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
TSMC to Remain Sole iPhone Chip Supplier, Could Provide ARM-Based Chips for Mac in 2020, Apple Car Chips in 2023
- Thread starter MacRumors
- Start date
- Sort by reaction score
A new 7nm fab is about $12 billion.
Developing the latest Intel processors is probably $2.5 billion.
Throw in $300-$500 million for mask sets at 7nm.
That’s a $15 Billion investment to get the first products out the door. You’d better be sure you’re going to sell a boat load of them.
I’m not convinced we will see too much at 7nm.
Apple doesn’t need the headache of owning somebody like TSMC, some of whose large customers of the 465 or so they currently have, would be very concerned about TSMC being bought by Apple - enough perhaps to go elsewhere.
And I’ve no idea what the anti trust approvals & other government approvals would result in.
Developing the latest Intel processors is probably $2.5 billion.
Throw in $300-$500 million for mask sets at 7nm.
That’s a $15 Billion investment to get the first products out the door. You’d better be sure you’re going to sell a boat load of them.
I’m not convinced we will see too much at 7nm.
Apple doesn’t need the headache of owning somebody like TSMC, some of whose large customers of the 465 or so they currently have, would be very concerned about TSMC being bought by Apple - enough perhaps to go elsewhere.
And I’ve no idea what the anti trust approvals & other government approvals would result in.
Depends on the app. But newer apps, especially ones written in Swift and Metal, but even stuff in contemporary C can mostly be recompiled and re-tested (if the vendor is still in business). There may even be a way, using bitcode, for Apple to do most of the recompile. The stuff written with a lot x86 asm is already dead or dying with the pulling of 32-bit support.
Apple has enough analytics to know exactly that percentage. If Apple switches, those bootcamp and enterprise VM users weren't really a significant share of Apple's market. They may make a lot of noise here, but very possibly don't spend the Billions needed for Apple to continue to support them across the product line (or maybe with just one expensive Pro box or something).
Apple's current processors don't support all those things a) yet or b) because they're mobile processors. It is highly likely that Apple will be supporting current desktop style components in their desktop class CPU's, if they eventuate.
And it wouldn't even have to be ARM-based. My bet is it will be an Apple full-custom design. There's more than just an adapting ARM-to-desktop move brewing.
[doublepost=1539821769][/doublepost]
Yeah, imagine if Apple moved faster developing a smartphone back when Motorola, Ericsson, and Nokia ruled. They could have been a contender.
Last edited:
Rumors have suggested Apple is planning to transition away from Intel chips to its own custom-made chips starting as early as 2020, which Kuo reiterates in today's report.
Let's hope this is not true. I'd drop Apple's whole eco-system if that happened.
The reason I have a Mac is because with VMWare's "Fusion" the Mac becomes a universal platform. Then the reason I have an iPhone and not Andriod is because I already have a Mac. But if the Mac turns into an iPad with a keyboard I'd move to Linux and Android.
My guess is that almost every Mac user would also "bail". I think a large fraction of Mac users bought the mac because it can run Windows apps at near-native speeds
[doublepost=1539829889][/doublepost]
It does not workthat way. A MacOS/ARM binary will not run on the Pi3.
My guess that anything you'd run on the Pi3 is open source so can you can compile it yourself.
I've been hedging my bets and moving my real-work off Apple equipmnt. I'm astounded at the much better price/performance ratio I see with Linux of PC hardware. I spend $500 on an off lease HP "Z" workstation with 64GB ECC RAM and a 16-core Xeon. This machine is 2 years old and blows away anything Apple has ever made.
I think Apple should buy HP gear and re-label it and sel it with Mac OS on it and we'd all be better off.
Is it a coincidence that after the announcement of Photoshop CC 2019 for iOS that this rumour is released.
Seems if any app that has a mobile version or a full port of its desktop equivalent that is iOS compliant may work with this ARM macOS entry level lineup.
ARM macOS:
1. MacBook
2. MacMini
3. iMac
x64 macOS:
1. MacBook Pro
2. Mac Pro
3. iMac Pro
Interesting development and Apple has the experience and history to pull this transition again. PowerPC > x86 > ARM/x64 > ARM
There is definitely one marked benefit for Apple to transition away from Intel chips and that is End of Hackintoshes.
I have a brain and have been observing apple for decades. It ain’t happening.
[doublepost=1539831599][/doublepost]
You must be young. You missed 68k->PowerPC.
Adding one more to the development history of Apple.
MOS 6502 > Motorola 68000-40 (CISC ) > Power PC > x86 > ARM-64/x64 > ARM (Apple Custom ISA)
Some people predict that Apple might switch to RISC-V Design at some point in Future. I dont see why Apple would do that other than to save royalty fees given per CPU to ARM Holdings.
6ps/mm interchip time of flight?
[doublepost=1539832013][/doublepost]
6502 to 68k doesn’t count since they didn’t really maintain compatibility?
I don’t think they will do RISC-v. They would likely still have to pay patent license fees to ARM, so not much point. I do think they’ll continue to morph the instruction set.
Um. Or people running OSX on their Raspberry Pi or other ARM SBCs....
[doublepost=1539832726][/doublepost]
Exactly. Here's that chart again folks of how LITTLE MacOS matters in the real world away from animoji
If only that were so easy people would be running iOS on their androids already, A series of chips have custom Apple ISA not found on any other ARM chip. Infact other than few parts of subsystem design from ARM many parts of the chip are designed by Apple and covered by patents.
[doublepost=1539832833][/doublepost]
Aah True but not entirely, Mac OS 9 Did run on both 68K and Power PC. OSX first ran on Intel.
Last edited:
True but I'm sure if it came to it the Hackintosh community life....would find a way
There are only 2 x86_64 bit instruction set vendors in the World. Why? Because they've cross-licensed and exclusive relationship and patented the rest.
They are AMD and Intel.
Apple isn't going to custom anything with the most common instruction set in the world for Desktop/Laptop/Enterprise Server markets.
By the way, all vendors shopping around to replace Intel in the Enterprise Server space are buying AMD EPYC solutions.
Having worked for Apple Engineering and Enterprise Services there is no shot ARM will enter other than being a custom security SoC for Servers.
AMD is the only option for Apple besides Intel on macOS.
not gonna happen, intel and amd holds all the patents to x64 and 86 related materials. apple is about 20 years late to the game if they wanna design their own x64 chips
Apple always does parallel transitions they would release an OS/Software update for both the new and the old and give a transition time to itself and its customers before successively moving and eventually completely transitioning to the new platform. Remember rosetta OS-7 68K to Power PC transition, both were supported till OS9. Even with OSX they did support PowerPC software through Rosetta Stone Dynamic Translation technology based Universal Apps, before completely moving to Intel Native Apps.
NeXTSTEP first ran on Motorola 68k, then HP-PA RISC, then Sun Sparc and eventual Intel x86. The only architecture that remains are Sun SPARC and x86, with only x86_64 being the instruction set Apple relies on.
SPARC is a dead architecture, thus we are down to x86_64, and no Apple isn't going to start from scratch as is evident by the fact they adopted a 15 year old architecture when it was well worn ala ARM being the choice in Mobile instead of MIPS.
ARM was already getting used in Mobile even at the time Apple released its iPhone. MIPS was popular in Supercomputing space.
Why not switch back to PowerPC? IBM is no longer a competitor either. My point is that this makes no sense. 
This would destroy everything Steve Jobs intended to do. Even the playing field.
That is, an Apple Mac can be a great computer, but if you demand mediocrity or your company does, you don't need to avoid Apple anymore and thus answering the most common question of Apple, marketshare.
If Tim Cook does this, Apple Macs will not cost any less. Forget that notion. Cost has never been their business model and never will be, well unless Tim Cook wants to go down the John Sculley route, which would end badly for him.
This would destroy everything Steve Jobs intended to do. Even the playing field.
That is, an Apple Mac can be a great computer, but if you demand mediocrity or your company does, you don't need to avoid Apple anymore and thus answering the most common question of Apple, marketshare.
If Tim Cook does this, Apple Macs will not cost any less. Forget that notion. Cost has never been their business model and never will be, well unless Tim Cook wants to go down the John Sculley route, which would end badly for him.
Yeah with programs that are already writ for OSX, there'll be a transition, but there's a ton of massive software with half a century of development behind it that won't be coming along for the ride. It's already a little hokey that we have to rely on virtualization, but once that's no longer an option, Apple will have walked away from our creative industries in total. If they want to bring designers, architects & engineers along with them, it'll take more than another Rosetta. They'd have to buy Dassault or Autodesk (Dassault plz) and squeeze mac-compatible versions out of them by force, because these are huge applications, and even at the best of times, the general feeling has been that Apple's updates are too fickle and unreliable to justify the expense of rewriting such massive software for their OS. Dumping Intel for ARM architecture would just be proving that point. They've already fallen well behind in hardware, aren't even in the pro tablet realm, but throwing compatibility out the window too, and providing nothing to fill in for that market segment would be... probably more acceptable for portables, but for desktops, an "interesting" choice.