Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
N5P:
  • 11" M2 iPad Pro tablet
  • 12.9" M2 iPad Pro tablet
  • 13" M2 MacBook Air laptop
  • 13" M2 MacBook Pro laptop
  • 24" M2 iMac AIO desktop
  • M2 Mac mini headless desktop
  • M2 Pro Mac mini headless desktop
  • 14" M2 Pro MacBook Pro laptop
  • 16" M2 Pro MacBook Pro laptop
N3:
  • 14" M2 Max MacBook Pro laptop
  • 16" M2 Max MacBook Pro laptop
  • M2 Max Mac Studio headless desktop
  • M2 Ultra Mac Studio headless desktop
  • 27" M2 Max iMac AIO desktop
  • 32" M2 Ultra iMac Pro AIO desktop
  • M2 Ultra Mac Pro headless desktop
  • M2 Extreme Mac Pro headless desktop
 
N5P:
  • 11" M2 iPad Pro tablet
  • 12.9" M2 iPad Pro tablet
  • 13" M2 MacBook Air laptop
  • 13" M2 MacBook Pro laptop
  • 24" M2 iMac AIO desktop
  • M2 Mac mini headless desktop
  • M2 Pro Mac mini headless desktop
  • 14" M2 Pro MacBook Pro laptop
  • 16" M2 Pro MacBook Pro laptop
N3:
  • 14" M2 Max MacBook Pro laptop
  • 16" M2 Max MacBook Pro laptop
  • M2 Max Mac Studio headless desktop
  • M2 Ultra Mac Studio headless desktop
  • 27" M2 Max iMac AIO desktop
  • 32" M2 Ultra iMac Pro AIO desktop
  • M2 Ultra Mac Pro headless desktop
  • M2 Extreme Mac Pro headless desktop
That progression is kinda logical except that no timeline is included.

1) It would be nuts (and a first) for Apple to go more than a year without an update to MBPs; especially since M2 in the lower end laptops are on the street now, cheaper stronger Mac Studios have now been out since March, and Apple has built a nice halo around the M2. Every day Apple is missing sales by not having high end M2 MBPs. That will be driving management crazy.

2) Less importantly, I suspect that Mac Pro may present differently.
 
As well as any delay in shipping devices that demonstrate Apple Silicon running on N3 chips... Because that flow from TSMC will probably start in early 2023.
Apple can easily deliver M2 MBPs made with good-yield n5 chips in 2022, so I think that they will. Delaying until 2023 to wait for poor-yield (therefore costly) n3 chips makes no sense.
 
  • Like
Reactions: motrek and throAU
Apple can easily deliver M2 MBPs made with good-yield n5 chips in 2022, so I think that they will. Delaying until 2023 to wait for poor-yield (therefore costly) n3 chips makes no sense.

Exactly. And N5P will still outperform the rest of the market in terms of power:watt and increase their profit margin as they won't be paying TSMC per wafer for a much larger number of defective parts.
 
  • Like
Reactions: Tagbert
LOL... no.

Apple isn't going to do a 5nm M2-Pro/Max, and then a 3nm M2-Pro/Max
Why not? Anyone with a scintilla of sense is waiting to see what M2 MBPs look like before buying now-old M1 MBP tech. Today Apple can resurrect high end MBP sales by introducing nicely evolved M2 MBPs with 5nm chips - - and make good profit using M2s off a high-yield 5nm line. Even if 3 nm can reach reasonable production levels soon, the 3nm is new enough that yields will be inferior for weeks, meaning 3nm boxes would cost Apple substantially more.

And buyers like me are waiting for the M2 generational improvements and box improvements (WiFi and the like). We are not waiting for the latest 3nm technology because the 5nm tech has already proven to be plenty powerful and energy efficient. Folks like me will buy immediately upon release of M2 MBPs.

Then in 2023 when 3nm yields have improved Apple can wow us with the 3nm gains. And keep pricing down and minimize delivery delays.
 
  • Like
Reactions: rcappo
Why not? Anyone with a scintilla of sense is waiting to see what M2 MBPs look like before buying now-old M1 MBP tech. Today Apple can resurrect high end MBP sales by introducing nicely evolved M2 MBPs with 5nm chips - - and make good profit using M2s off a high-yield 5nm line. Even if 3 nm can reach reasonable production levels soon, the 3nm is new enough that yields will be inferior for weeks, meaning 3nm boxes would cost Apple substantially more.

And buyers like me are waiting for the M2 generational improvements and box improvements (WiFi and the like). We are not waiting for the latest 3nm technology because the 5nm tech has already proven to be plenty powerful and energy efficient. Folks like me will buy immediately upon release of M2 MBPs.

Then in 2023 when 3nm yields have improved Apple can wow us with the 3nm gains. And keep pricing down and minimize delivery delays.

I think the point the OP was making is that a "3nm M2" would be significantly better than a 5nm M2, and Apple wouldn't pass on the opportunity to name it something else and otherwise advertise it to customers. "M3" is an option.
 
  • Like
Reactions: rcappo and Tagbert
I think the point the OP was making is that a "3nm M2" would be significantly better than a 5nm M2, and Apple wouldn't pass on the opportunity to name it something else and otherwise advertise it to customers. "M3" is an option.
Ahh. Thanks for that, makes sense. Apple is unlikely to treat the huge 5nm to 3nm evolution as simply a speed bump in marketing. Nor should they, since the 3nm process will undoubtedly allow major gains.
 
Last edited:
Why not? Anyone with a scintilla of sense is waiting to see what M2 MBPs look like before buying now-old M1 MBP tech. Today Apple can resurrect high end MBP sales by introducing nicely evolved M2 MBPs with 5nm chips - - and make good profit using M2s off a high-yield 5nm line. Even if 3 nm can reach reasonable production levels soon, the 3nm is new enough that yields will be inferior for weeks, meaning 3nm boxes would cost Apple substantially more.

And buyers like me are waiting for the M2 generational improvements and box improvements (WiFi and the like). We are not waiting for the latest 3nm technology because the 5nm tech has already proven to be plenty powerful and energy efficient. Folks like me will buy immediately upon release of M2 MBPs.

Then in 2023 when 3nm yields have improved Apple can wow us with the 3nm gains. And keep pricing down and minimize delivery delays.
If Apple releases 5nm M2-Pro/Max then the 3nm will be reserved for the M3-Pro/Max.
 
  • Like
Reactions: rcappo and Tagbert
Yeah, it would be weird, for sure. I could see 3nm being for M3, starting with mobile (fall 2023 release)—I mean, that's typically where they start with newer/thinner chips. I'd still consider an M2 MBP in the spring of 2023 though...
Well, M3 at N3 would be nice, although it'd be best to have the next Mx Pro chips already at 3nm given that the M2 is hotter chip on its own.
 
Interesting thoughts/quotes from this article:


The N3 node is weird. It is TSMC’s first 3-nanometer class process technology, but it won’t be the mainstream node everyone will utilize.

N3E differs significantly enough from N3 to be treated as an entirely different migration path from N5. TSMC also makes its comparisons for N3E against N5 and not N3 for this reason.

Makes you wonder: Maybe Apple uses:
- N3 to push Mac Pro out the door after the end of the year
- N5P for M2 Pro/Max for current year replacements (as Ming-Chi Kuo predicts)
- N3E for M3 spectrum of chips late next year for 3nm goodness
 
  • Like
Reactions: Tagbert
Commercial Times reiterates 3nm for M2 Pro and M2 Max.

 
  • Like
Reactions: theorist9
Commercial Times reiterates 3nm for M2 Pro and M2 Max.

If the M2 Pro/Max are on 3nm, then I'll be pleasantly surprised if the M2 Pro/Max MBP's are released fall 2022—I thought we weren't supposed to be seeing 3 nm devices until spring 2023.
 
If the M2 Pro/Max are on 3nm, then I'll be pleasantly surprised if the M2 Pro/Max MBP's are released fall 2022—I thought we weren't supposed to be seeing 3 nm devices until spring 2023.
Yes, Commercial Times has repeatedly stated N3 for M2 Pro/Max and today just reiterated that, but Ming-Chi Kuo said M2 Pro/Max MacBook Pros are coming soon and thus are too early for N3.

I personally believe M2 Pro and M2 Max are indeed coming soon, but the N3 part has left me confused. There are a number of factors here to digest though. Here's how I look at it:

1. A lot of people have said that most high volume N3 family will be on N3E, but it's far too early for N3E, so that's out. N3E would likely be for products in the second half of 2023 or later. So, Apple will likely make use of N3E for M3, but not for M2 Pro/Max.

2. Perhaps Apple did in fact contract TSMC for its early adopter N3 process (instead of waiting for N3E) but TSMC has specifically said that they will see no N3 revenue until early 2023. Either TSMC is lying / being too conservative with their dates, or else it means actual Macs with those N3 chips wouldn't see the light of day until Q1 2023.

3. Despite the fact that TSMC said they won't get N3 revenue until 2023, their own roadmaps state that high volume manufacturing on N3 begins/began in the second half of 2022.

4. Ming-Chi Kuo seems fairly convinced that based on supply chain info, M2 Pro/Max MacBook Pros are being made now/soon. Factoring in TSMC's claim that no N3 revenue will come until 2023, he's taken that to mean M2 Pro/Max are not N3.

5. @Amethyst states that the new Mac minis and new 14/16" MacBook Pros are ready to go. The Mac minis may not necessarily factor into this discussion, because they could be just M2, but 14/16" MacBook Pros will have M2 Pro/Max. This supports Ming-Chi Kuo's contention that the new MacBook Pros are seeing production this quarter.

In summary, it looks like M2 Pro/Max Macs are likely already in production, but we have conflicting info as to whether they're going to be N3 or not.

tsmcroadmap.jpeg


BTW, if you're wondering what the big deal about N3 vs N3E is, here is a table that summarizes it. Essentially, N3 is good, but N3E is better.


Screen Shot 2022-10-25 at 2.01.26 PM.png
 
Last edited:
Yes, Commercial Times has repeatedly stated N3 for M2 Pro/Max and today just reiterated that, but Ming-Chi Kuo said M2 Pro/Max MacBook Pros are coming soon and thus are too early for N3.

I personally believe M2 Pro and M2 Max are indeed coming soon, but the N3 part has left me confused. There are a number of factors here to digest though. Here's how I look at it:

1. A lot of people have said that most high volume N3 family will be on N3E, but it's far too early for N3E, so that's out. N3E would likely be for products in the second half of 2023 or later. So, Apple will likely make use of N3E for M3, but not for M2 Pro/Max.

2. Perhaps Apple did in fact contract TSMC for its early adopter N3 process (instead of waiting for N3E) but TSMC has specifically said that they will see no N3 revenue until early 2023. Either TSMC is lying / being too conservative with their dates, or else it means actual Macs with those N3 chips wouldn't see the light of day until Q1 2023.

3. Despite the fact that TSMC said they won't get N3 revenue until 2023, their own roadmaps state that high volume manufacturing on N3 begins/began in the second half of 2022.

4. Ming-Chi Kuo seems fairly convinced that based on supply chain info, M2 Pro/Max MacBook Pros are being made now/soon. Factoring in TSMC's claim that no N3 revenue will come until 2023, he's taken that to mean M2 Pro/Max are not N3.

5. @Amethyst states that the new Mac minis and new 14/16" MacBook Pros are ready to go. The Mac minis may not necessarily factor into this discussion, because they could be just M2, but 14/16" MacBook Pros will have M2 Pro/Max. This supports Ming-Chi Kuo's contention that the new MacBook Pros are seeing production this quarter.

In summary, it looks like M2 Pro/Max Macs are likely already in production, but we have conflicting info as to whether they're going to be N3 or not.

View attachment 2101845View attachment 2101846
Or we could get 5 nm M2 MBP's in the fall, followed by a spec bump to 3 nm in the spring.

I think it would be cool if Apple went back to their old practice of doing frequent (twice-yearly) spec bumps. Though that could be more costly for them now, since then they were just popping in Intel's new CPU's, while now they'd need to keep updating the CPU's themselves.

OTOH, if they're going to be relasing 3 nm M2 Pro/Max CPU's anyways for other devices in the spring, then updating the MBPs' to those wouldn't be any more costly than when they were using Intel.
 
Or we could get 5 nm M2 MBP's in the fall, followed by a spec bump to 3 nm in the spring.

I think it would be cool if Apple went back to their old practice of doing frequent (twice-yearly) spec bumps. Though that could be more costly for them now, since then they were just popping in Intel's new CPU's, while now they'd need to keep updating the CPU's themselves.

OTOH, if they're going to be relasing 3 nm M2 Pro/Max CPU's anyways for other devices in the spring, then updating the MBPs' to those wouldn't be any more costly than when they were using Intel.
Ugh... people here are already whining about being "forced to upgrade" if Apple brings out MBP with M2 X chips a year after releasing them with M1 X chips. The cries will be deafening if Apple were do do more than one spec bump in a year. :oops:
 
  • Haha
Reactions: EugW
Ugh... people here are already whining about being "forced to upgrade" if Apple brings out MBP with M2 X chips a year after releasing them with M1 X chips. The cries will be deafening if Apple were do do more than one spec bump in a year. :oops:
Yes, that's true :D. But, as you know, all consumers benefit from more frequent spec bumps. The loss from a new machine coming out just a few months after you bought yours is mainly psychological. And the solution for those who feel bad because "gee, if I'd only known I could have just waited a few months" would be for Apple to do spec bumps more regularly. Then if you're 3 months into a current release, you know there's a reasonable possibility that there's going to be yet another release 3 months hence.
 
Last edited:
Sorry, not following--process changes do change the performance.
They changed the process, but kept the design and the clock speed the same, so performance in benchmarks and applications was exactly the same. The chip name didn't change either. However, the chips would have used less power on the new process.

Also, for another chip they had two different companies making the same chip on two different processes at the same time. So you could get an iPhone based on one process or the other. No way of knowing.
 
They changed the process, but kept the design and the clock speed the same, so performance in benchmarks and applications was exactly the same. The chip name didn't change either. However, the chips would have used less power on the new process.
OK, but that's certainly the least likely approach Apple would take if they were to switch the 14"/16" MBPs from 5 nm M2 to 3 nm M2—instead, they'd keep the thermal envelope (about) the same, and use all (or nearly all) of the process improvement for increased performance. It seems extraordinary unlikely that they'd instead use it all to increase efficiency, keeping performance essentially unchanged.

Also, for another chip they had two different companies making the same chip on two different processes at the same time. So you could get an iPhone based on one process or the other. No way of knowing.
I think that was when they switched manufacturers for one of the iPhone A chips. Regardless, that's not applicable here, since we're discussing a change from TSMC 5 nm to TSMC 3 nm.
 
OK, but that's certainly the least likely approach Apple would take if they were to switch the 14"/16" MBPs from 5 nm M2 to 3 nm M2—instead, they'd keep the thermal envelope (about) the same, and use all (or nearly all) of the process improvement for increased performance. It seems extraordinary unlikely that they'd instead use it all to increase efficiency, keeping performance essentially unchanged.


I think that was when they switched manufacturers for one of the iPhone A chips. Regardless, that's not applicable here, since we're discussing a change from TSMC 5 nm to TSMC 3 nm.
Bottom line, I don't think that's going to happen. Whatever it gets manufactured on, it's gonna stay on it.

BTW, apparently mass production at 3nm has not yet started.


Now I will talk about our N3 and N3E status. Our N3 is on track for volume production later this quarter with good year. We expect a smooth ramp in 2023, driven by both HPC and smartphone applications. Our customers' demand for N3 exceeds our ability to supply partially due to the ongoing tool delivery issues, and we expect N3 to be fully utilized in 2023.
 
Or we could get 5 nm M2 MBP's in the fall, followed by a spec bump to 3 nm in the spring.
That seems most likely to me, because of yields. The 5nm and the 4nm processes are well established and therefore should have good yields and lower cost while the 3nm process is very very new and therefore will have poor yields and consequently very high cost well into 2023. I would guess any 3nm spec bump as Q3, not spring.

IMO many of the folks like me waiting for M2 are primarily waiting for the v2 generational evolution from Gen 1, not for new super-hot performance. We will be fine with smooth, well established 5nm M2 Pro/Max in our new MBPs, especially if Apple throws us a bone or two like lower prices, WiFi 6E, etc.

New Mac Pros however IMO must get new super-hot performance, which probably means N3 to get larger transistor count. How Apple deals with (sooner) N3P vs. (later, better) N3E is a big question.
 
  • Like
Reactions: theorist9
New Mac Pros however IMO must get new super-hot performance, which probably means N3 to get larger transistor count. How Apple deals with (sooner) N3P vs. (later, better) N3E is a big question.
I was thinking the same thing--that it would be best if they waited to release the AS Mac Pro until they had it on a 3 nm process, which would result in enough of a die shrink for them to have hardware RT, which takes up a lot of space (and that would also mean M3 instead of M2). With the low volume of the Mac Pro, they'd probably be able to get enough 3 nm chips for a spring 2023 release. But I don't know if their microarchitecture development is far enough along for them to be able to release an M3 Mac Pro with hardware RT at that time.
 
  • Like
Reactions: Allen_Wentz
I was thinking the same thing--that it would be best if they waited to release the AS Mac Pro until they had it on a 3 nm process, which would result in enough of a die shrink for them to have hardware RT, which takes up a lot of space (and that would also mean M3 instead of M2). With the low volume of the Mac Pro, they'd probably be able to get enough 3 nm chips for a spring 2023 release. But I don't know if their microarchitecture development is far enough along for them to be able to release an M3 Mac Pro with hardware RT at that time.
Facilitating hardware ray tracing in the Mac Pro with Apple SoC would be exactly the kind of huge advance Apple needs to make an appropriate statement. And my expectation is that once it works in MP then it would also work (albeit more slowly) on future N3E MBPs and Studios.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.