Apple Silicon Timeline (Gurman)

[Jorbanead]

In the latest MacRumors Podcast, Mark Gurman discussed the original timeline for Apple Silicon. He stated the M1 Pro/Max notebooks were supposed to come in the Spring of 2021, and possibly M2 macs in the fall of 2021. Essentially on an annual cadence, with M1/M2/M3 in the fall and the Pro/Max variants mid-cycle in the spring.

We all know it didn’t pan out this way. Gurman suggests Apple delayed the MacBooks Pro not once, but twice. It seems based on current rumors that we may not get M2 macs until this fall.

What do you think Apple is going to do going forward? If Gurman is correct, this means we are essentially 1 year behind the original schedule. I would assume the actual chip team isn’t behind schedule though as we got the A15 this year, meaning they likely have M2 already done. Rumors suggested delays were due to other issues like miniLED.

With the latest AMD and Intel chips as well, Apple also has a bit of catching up to do (not in performance per watt, but single core performance). Do you think Apple will eventually skip a generation (but still call it M2?) to make up for the delays?

 

[leman]

In the latest MacRumors Podcast, Mark Gurman discussed the original timeline for Apple Silicon. He stated the M1 Pro/Max notebooks were supposed to come in the Spring of 2021, and possibly M2 macs in the fall of 2021. Essentially on an annual cadence, with M1/M2/M3 in the fall and the Pro/Max variants mid-cycle in the spring.

Sounds very reasonable.

We all know it didn’t pan out this way. Gurman suggests Apple delayed the MacBooks Pro not once, but twice. It seems based on current rumors that we may not get M2 macs until this fall.

I thought the latest sources said M2 Air in spring?

With the latest AMD and Intel chips as well, Apple also has a bit of catching up to do (not in performance per watt, but single core performance). Do you think Apple will eventually skip a generation (but still call it M2?) to make up for the delays?

That might be a reasonable thing to do, but also kind of depends on how the delays will work out. If they have to delay the Air until fall, then releasing M2 as A16-based arch might make sense.

 

[JMacHack]

In the latest MacRumors Podcast, Mark Gurman discussed the original timeline for Apple Silicon. He stated the M1 Pro/Max notebooks were supposed to come in the Spring of 2021, and possibly M2 macs in the fall of 2021. Essentially on an annual cadence, with M1/M2/M3 in the fall and the Pro/Max variants mid-cycle in the spring.

We all know it didn’t pan out this way. Gurman suggests Apple delayed the MacBooks Pro not once, but twice. It seems based on current rumors that we may not get M2 macs until this fall.

What do you think Apple is going to do going forward? If Gurman is correct, this means we are essentially 1 year behind the original schedule. I would assume the actual chip team isn’t behind schedule though as we got the A15 this year, meaning they likely have M2 already done. Rumors suggested delays were due to other issues like miniLED.

With the latest AMD and Intel chips as well, Apple also has a bit of catching up to do (not in performance per watt, but single core performance). Do you think Apple will eventually skip a generation (but still call it M2?) to make up for the delays?

1. Who says they had to do anythingthey weren’t already gonna do? They weren’t there only company set back by covid.

2. Performance per watt is pretty much the reason for Apple Silicon. The benchwank was a nice cherry on top.

3. ASi has other performance advantages other than single thread (floating point was off the charts, and fp is a big deal in browser and web apps.)

4. The friggin processors aren’t in peoples hands yet. Quit fretting.

 

[Jorbanead]

1. Who says they had to do anythingthey weren’t already gonna do? They weren’t there only company set back by covid.

2. Performance per watt is pretty much the reason for Apple Silicon. The benchwank was a nice cherry on top.

3. ASi has other performance advantages other than single thread (floating point was off the charts, and fp is a big deal in browser and web apps.)

4. The friggin processors aren’t in peoples hands yet. Quit fretting.

You seem to mistake this post as another doom and gloom. I don’t think Apple is in any danger here. I’m simply just posing a question for discussion as there are some smart folks in these forums that understand how this works more than I do.

While I understand performance per watt is the main goal, and I also very much understand ASi involves additional factors outside of SC, like encode/decode, NPU blocks, accelerators etc. Apple still very much cares about CPU performance and it would be foolish to think they wouldn’t want to be at the top. They boast about it at every single iPhone keynote, and they hailed M1 as being the fastest chip in a lightweight notebook.

Reason why this matters is people compare apple chips to Intel/AMD/Nvidia all the time. They need to stay competitive if they want to charge the prices they do.

Yes the processors aren’t out yet: but half of what we do on these forums is discuss things that haven’t happened yet. Rumors that are… Mac… related.

 

[cupcakes2000]

there are some smart folks in these forums that understand how this works more than I do.

Well - maybe there are. But there are also a bunch of loudmouths that know nothing about anything in these forums, especially when compared with other forums. It’s not the fault of macrumours, more the fault of their popularity.
They shout louder. That’s the problem I think. It’s akin to going to Reddit for your sole news source. It’s more about clicks and likes than it is about knowledge.
That aside - when dealing with rumours one has to consider exactly that - they’re precisely rumours.
I think that Apple are probably experiencing the same covid related issues as their competitors- but also are still publicly within their stated timeline. I’m not sure any issues exist until that ceases to be.

 

[ian87w]

I believe the component shortages is dampening Apple’s plans significantly. I’ve been wondering also why Apple seems to be taking its sweet time in transitioning to its own silicon. You would think they will want to get rid intel ASAP and save the cost of making those remaining intel models.

Seeing the M1 MacBooks went for a year without refreshes seem a telltale something is delaying Apple.

As for intel/AMD gaining, yes, but it will also take time for actual PCs with those new chips to be shipping. Right now, 11th gen intel and Ryzen5 machines are just getting more commonplace. By the time 12th gen are common and shipping, Apple would have M2 MacBooks shipping as well. Everybody is being bottlenecked by the components shortages.

 

[robco74]

Apple has released plain A chips yearly with each iPhone release, but the bumped up X variants used in the iPad Pro were two years or more between updates. And of course, while Apple designs the chips, they have to rely on TSMC to actually produce them. And as you pointed out, there are shortages of other components as well. I wouldn't be too surprised if Apple announced a delay at the next WWDC.

I am personally interested to see how the new AMD 6xxxU chips fare against M1 in the ultraportable space now that the iGPU has been updated. We saw this sort of back and forth in the early PPC days. As long as Apple remains reasonably competitive, I don't think they need to worry too much. Raw performance is nice, but can only speed up workflows so much.

I don't think we need to worry about a return to the days when the PowerBook was stuck on the G4 was getting royally spanked by Intel.

 

[JPack]

There was zero chance of an annual cadence for M-series. Launching M1 iMac in April 2021 told everyone what Apple's plan was. It doesn't use mini LED nor any new breakthrough technologies. Yet, it was launched 6 months after M1. Does anybody believe Apple had to delay the M1 iMac?

As early as mid-2020, Kuo predicted MacBook Pro 14 and 16 with mini LED would come in late 2Q21 or 3Q21. Apple didn't miss their timelines. This is one of those cases where Mark Gurman missed.

Kuo: Apple Silicon Macs to Include 13-inch MacBook Pro and MacBook Air This Year, 14.1-inch and 16-inch MacBook Pro Models Next Year

At last month's WWDC, Apple officially announced that its Mac computers will be transitioned from Intel x86 to homegrown Apple Silicon chips....

www.macrumors.com

 

[Boil]

There was zero chance of an annual cadence for M-series. Launching M1 iMac in April 2021 told everyone what Apple's plan was. It doesn't use mini LED nor any new breakthrough technologies. Yet, it was launched 6 months after M1. Does anybody believe Apple had to delay the M1 iMac?

Delay the M1 24" iMac, no; but the production of a brand new chassis might slow things down a bit...?

 

[julesme]

[…] Do you think Apple will eventually skip a generation (but still call it M2?) to make up for the delays?

I think the answer is probably, “yes,” there could be a scenario where the M series skips a generation to utilize the latest fab process. My logic is that the iPhone is reliably on a 12 month cycle and (in my opinion) the M series is more likely to be 18 months, given how often iPads, MBPs, and iMacs are historically refreshed. So just based on that differential, the M series will eventually need to skip a generation or it will fall behind.

I’m curious to hear people’s thoughts about whether they think M series is ultimately better suited to 12 months or something less frequent, like 18 months?

 

[Boil]

I think the answer is probably, “yes,” there could be a scenario where the M series skips a generation to utilize the latest fab process. My logic is that the iPhone is reliably on a 12 month cycle and (in my opinion) the M series is more likely to be 18 months, given how often iPads, MBPs, and iMacs are historically refreshed. So just based on that differential, the M series will eventually need to skip a generation or it will fall behind.

I’m curious to hear people’s thoughts about whether they think M series is ultimately better suited to 12 months or something less frequent, like 18 months?

Had to dig a bit to find it, so many fluff news articles the last month or so...!

Apple silicon roadmap based on 18 month upgrade cycle...

 

[Jorbanead]

Had to dig a bit to find it, so many fluff news articles the last month or so...!

Apple silicon roadmap based on 18 month upgrade cycle...

I saw this. It’s hard to know if they have inside sources that are saying this is the preferred plan going forward, or if this is the plan right now due to the pandemic - but it’s not what apple intends to do once things are back to normal.

I wouldn’t be surprised if the M# series chips are annual, while the higher-end Mac Pro model chips are on an 18-month cycle. There’s way more volume on the consumer side, while professionals are less likely to upgrade often. There’s just less demand.

 

[Jorbanead]

, it was launched 6 months after M1. Does anybody believe Apple had to delay the M1 iMac?

sure. ID was occupied by other things. Pandemic related shortages. Apple marketing wanted to focus on entry-level notebooks and the mini for devs. I could think of several reasons why they’d delay the iMac.

M series is more likely to be 18 months, given how often iPads, MBPs, and iMacs are historically refreshed.

Thing is that now there is way more volume of devices that will utilize these chips. Before it was just the iPad Pro that had its own custom chip separate from the baseline A series, but now that category of chip is also being used in 4 additional products. So apple has more incentive to do annual update cycles. Historically, consumer macs have often been on annual upgrade cycles.

 

[JPack]

Delay the M1 24" iMac, no; but the production of a brand new chassis might slow things down a bit...?

If you look at the teardown for M1 iMac or MacBook Pro, there's no indication the chassis is complex. If anything, the M1 chassis is a lot simpler than the Intel-based iMac.

 

[JPack]

sure. ID was occupied by other things. Pandemic related shortages. Apple marketing wanted to focus on entry-level notebooks and the mini for devs. I could think of several reasons why they’d delay the iMac.

Nobody expected an Apple Silicon iMac to launch in 2020. The Intel-based iMac had only been refreshed August 2020.

It was 8 months between MBA (Ice Lake) and MBA (M1). We saw that same 8 month gap between the last Intel iMac update (August 2020) and M1 iMac (April 2021).

In short, there is no indication any of the Apple Silicon Macs were delayed. Mac and iPad combined sells at least an order of magnitude fewer units than iPhone. It's silly to think the M-series chips would have the same cadence as iPhone. Silicon design costs are doubling each major node. $300 million at 7nm. $500 million at 5nm. $1 billion at 3nm. The market for Mac and iPad aren't growing at the same pace and neither are the margins.

 

[Jorbanead]

Nobody expected an Apple Silicon iMac to launch in 2020. The Intel-based iMac had only been refreshed August 2020.

That Intel iMac was only the 27” while the 21.5” iMac had not been updated for two years.

Gurman is the most accurate leaker to date based on volume, so I’m more inclined to believe what he says. Especially since this also coincides with a few other rumors that we were supposed to get the MacBook Pro’s at WWDC this year.

Silicon design costs are doubling each major node. $300 million at 7nm. $500 million at 5nm. $1 billion at 3nm.

Silicon design costs? Apple isn’t doubling their design team each year. That argument doesn’t make any sense. Are you referring to TSMC fab costs?

 

[robco74]

Apple sells far more iOS devices than they do Macs. They sell far more iPhones than iPads. Due to sheer volume, it's feasible to update the the A-series chip and internals for the iPhone each year, and refresh the design every other year or so. For iPads and Macs, not so much. It takes a longer to sell enough iPads and Macs to offset the initial costs incurred getting them into mass production. ROI for the iPhone happens much more quickly.

Also, the smartphone market is much more competitive. It's worth the cost for Apple to update the iPhone each year. But the iPad has little real competition in the tablet space. AMD and Intel don't usually make major updates each year, mostly just new SKUs with minor changes, so less pressure to develop new M-series chips each year.

 

[satcomer]

The things that slowed down Mac was the virus lockdowns in California! Employees of Apple can't go to the Apple Spaceship campus yet!

 

[Andropov]

Apple sells far more iOS devices than they do Macs. They sell far more iPhones than iPads. Due to sheer volume, it's feasible to update the the A-series chip and internals for the iPhone each year, and refresh the design every other year or so. For iPads and Macs, not so much. It takes a longer to sell enough iPads and Macs to offset the initial costs incurred getting them into mass production. ROI for the iPhone happens much more quickly.

It's true that they sell a lot more iPhones, but that doesn't necessarily mean that they can't release M and AX chips every year too. It's not like the ROI for iPhone SoC designs is of exactly one year. For all we know, the costs could be offset the first month after release for iPhones and the third for iPads or Macs.

Also, the phone/tablet SoC space is less competitive for them now, since they're years ahead of the competition. They can afford not updating the iPad SoC in two to three years and still have the fastest tablet SoC on the market. On laptops/desktops, not so much.

 

[EntropyQ3]

Nobody expected an Apple Silicon iMac to launch in 2020. The Intel-based iMac had only been refreshed August 2020.

It was 8 months between MBA (Ice Lake) and MBA (M1). We saw that same 8 month gap between the last Intel iMac update (August 2020) and M1 iMac (April 2021).

In short, there is no indication any of the Apple Silicon Macs were delayed. Mac and iPad combined sells at least an order of magnitude fewer units than iPhone. It's silly to think the M-series chips would have the same cadence as iPhone. Silicon design costs are doubling each major node. $300 million at 7nm. $500 million at 5nm. $1 billion at 3nm. The market for Mac and iPad aren't growing at the same pace and neither are the margins.

Those numbers aren’t anywhere near true for derivative designs. Design costs for the A14, M1, M1Pro and M1Max chips are NOT the same respectively. While there are devils hiding in the details (quite a few and strong ones too), it’s new functional units and connectivity that costs.
To try to clarify, if you already have a completely functional GPU core with cache and data interfaces on the process you want to use, putting in more of them is much MUCH easier than designing and debugging a GPU core in the first place. Hence why you tend to see "scaled" CPUs and GPUs - it’s comparatively cheap, and a way to leverage an existing design to adress a wider array of market segments.

That said, you do have a strong point in that leading edge silicon lithography costs favour high volume products! But it also favours keeping all your silicon similar, so if Apple has new cores on, say, TSMC 3nm for their phones, it makes financial sense for them to leverage those designs across all their devices. That is not to say that every new phone SoC will have a corresponding Mac SoC associated with it. It may make sense for Apple to be a bit more judicious in their choices of what and when to scale up for the iPad Pros and Macs.

 

[Bug-Creator]

As early as mid-2020, Kuo predicted MacBook Pro 14 and 16 with mini LED would come in late 2Q21 or 3Q21. Apple didn't miss their timelines. This is one of those cases where Mark Gurman missed.

One doesn't (fully) rule out the other. If the original "M1x" concept failed it did so in late 2020 which would have given Apple time and opportunity to change plans on what gets a redesign when.
The weird 1fan vs 2fan iMac versions strongly suggest that something is off.

My logic is that the iPhone is reliably on a 12 month cycle and (in my opinion) the M series is more likely to be 18 months,

18months would either mean that they:
- sell iPhones that have 6 months old cores
- sell Macs that have 6 months old cores
- do in between versions of these cores

I don't think that any plan they might have is set in stone and it will take a few generations to settle to a fixed schedule (if at all).

 

[mcnallym]

Don’t forget that unlike the Mac and iPad’s the iPhones are sold for several years.

Example being that on the Apple store today you can buy new

iPhone 11, 12 and 13 models, as well as the SE.

When a new iPad launches previous one removed from sale apart from refurbs. Same for Mac. Once the M2 mini or iPad Pro launches then the M1 versions will be withdrawn from sale.

So the iPhone soc around for much longer to get the ROI.

 

[boak]

One doesn't (fully) rule out the other. If the original "M1x" concept failed it did so in late 2020 which would have given Apple time and opportunity to change plans on what gets a redesign when.
The weird 1fan vs 2fan iMac versions strongly suggest that something is off.



18months would either mean that they:
- sell iPhones that have 6 months old cores
- sell Macs that have 6 months old cores
- do in between versions of these cores

I don't think that any plan they might have is set in stone and it will take a few generations to settle to a fixed schedule (if at all).

The iPad Air had higher single core performance than the iPad Pro in 2020. Not unprecedented.

 

[Bug-Creator]

@mcnallym

Plenty of Intel Mac that were sold as new while staying on an older generation of CPUs.
Base iPad is and has been for a long time based on an older A-series.

Wouldn't be surprised if the M1 Air and Mini stay around as bargain options after a M2 launch.

@boak

Both the Air and the Pro used the latest gen available at launch.

Also for iPhones and iPad Apple only competes against itself and being able to say "fastest xxxx ever" every 12 months was enough to sell them in millions.

With the Mac the have real competition, which does seem to entering a phase of quick advancement (both Intel and AMD).

 

[boak]

Mx Pro/Max are such low volume chips. Annual cadence makes little sense.