More evidence that Apple plans to update Apple Silicon yearly

[senttoschool]

"The M2 family was really now about maintaining that leadership position by pushing, again, to the limits of technology. We don't leave things on the table," says Millet. "We don't take a 20% bump and figure out how to spread it over three years...figure out how to eke out incremental gains. We take it all in one year; we just hit it really hard. That's not what happens in the rest of the industry or historically."

Apple execs on M2 chips, winning gamers and when to buy a Mac | TechCrunch

Apple's M series chips were incredibly well telegraphed when they arrived in late 2020. Apple had been designing its own silicon since the A4 appeared in

techcrunch.com

Other evidence/speculation include:

• Rumors were that M2 SoC was ready much earlier but was held back by the late new Air design.
• Reports and video file name suggest that the M2 Pro/Max were ready as early as October 2022, which would have made it one year since M1 Pro/Max
• M2 used A15 cores (not A16 in iPhone 14), which suggests Apple wants to take advantage of the iPhone SoC design update each year
• Reports are that M2 will be a short-lived generation because M3 is expected to launch on time
• Many speculators point to iPad SoCs, which were not updated yearly. However, the base M chips go into far more devices than old iPad SoCs. This makes it more economical to design and launch new SoCs yearly.
• M1 launched in the same quarter as A14, which suggests Apple develops the base M and the A series in tandem
• Apple has been fighting supply chain issues in the last 3 years, which caused many delays in new designs shipping.

Why is this important? Because gaining 10-20% in performance every two years is not impressive. But once a year? Now that's impressive and exciting.

 

[leman]

Absolutely, it’s pretty much a certainty that Apple is targeting a yearly upgrade schedule. Right now we have a delay of roughly a year. I’m quite sure that M3 family was originally supposed to arrive last fall.

 

[senttoschool]

Absolutely, it’s pretty much a certainty that Apple is targeting a yearly upgrade schedule. Right now we have a delay of roughly a year. I’m quite sure that M3 family was originally supposed to arrive last fall.

It makes you wonder what M3 will be. If it was supposed to launch last fall, it had to be made on 5nm. But because TSMC 3nm is now in volume production, it'd make more sense to use 3nm for M3. However, chips can't just change nodes suddenly. It takes years of planning.

This is why I think the original M3 using A16 is canceled. Instead, Apple will launch the M3 using A17, which was likely designed for 3nm from the start. If so, M3 is really M4. Or Apple always planned to skip the A16 cores.

Regardless, I think M3 will be the generation that will wow people again with 3nm and A17. Unless Apple gives us a dud and uses A16 and 5nm.

 

[leman]

It makes you wonder what M3 will be. If it was supposed to launch last fall, it had to be made on 5nm. But because TSMC 3nm is now in volume production, it'd make more sense to use 3nm for M3. However, chips can't just change nodes suddenly. It takes years of planning.

This is why I think the original M3 using A16 is canceled. Instead, Apple will launch the M3 using A17, which was likely designed for 3nm from the start. If so, M3 is really M4. Or Apple always planned to skip the A16 cores.

Regardless, I think M3 will be the generation that will wow people again with 3nm and A17. Unless Apple gives us a dud and uses A16 and 5nm.

No, 3nm was originally planned for 2022 and was delayed by TSMC. I’m fairly sure that A16 and M3 were supposed to be 3nm designs with completely new CPU and GPU (we even have parent details on the GPU).

In the end, what Apple shipped as A16 is likely a placeholder, basically an A15 that has been slightly tweaked for higher clocks.

I think it all makes sense if you think about it. Apple was claiming a two year transition, but A14-derived designs (both M1 and M2) don’t scale too well to desktop use. For the Mac pro to make any sense whatsoever they need faster, more scalable tech. We know that they have been working on a new more scalable on-chip network, hardware raytracing and new memory protocols. I believe all this tech was supposed to land in late 2022 and set the scene for a Mac Pro.

In fact, this opens way to speculation that a 3nm Mac Pro might make it debut at WWDC. Let’s wait and see.

 

[senttoschool]

No, 3nm was originally planned for 2022 and was delayed by TSMC. I’m fairly sure that A16 and M3 were supposed to be 3nm designs with completely new CPU and GPU (we even have parent details on the GPU).

This could be it but it seems unlikely.

Sure, we have reports of 3nm being delayed. But Apple should have received that news years ago, probably way before the public ever got wind of the delay. So Apple probably had time to adjust. It seems unlikely that A16 was backported to 5nm from 3nm. It seems more likely that A16 was designed for 5nm from the start.

Are you referring to rumors that ray tracing did not make it to A16?

 

[leman]

This could be it but it seems unlikely.

Sure, we have reports of 3nm being delayed. But Apple should have received that news years ago, probably way before the public ever got wind of the delay. So Apple probably had time to adjust. It seems unlikely that A16 was backported to 5nm from 3nm. It seems more likely that A16 was designed for 5nm from the start.

Are you referring to rumors that ray tracing did not make it to A16?

I never said that A16 was back ported from 3nm. What I am saying that I suspect that A16 is not actually A16. I almost believe that original A16 was supposed to be a 3nm chip with all new internal design but since 3nm was delayed Apple has scrambled toughener another variant of A15 and called it "A16". The actual A16 (probably slightly iterated upon) will then come out as A17.

At least this is the theory that explains why there are so few changes in the A16 we got.

 

[dmccloud]

I never said that A16 was back ported from 3nm. What I am saying that I suspect that A16 is not actually A16. I almost believe that original A16 was supposed to be a 3nm chip with all new internal design but since 3nm was delayed Apple has scrambled toughener another variant of A15 and called it "A16". The actual A16 (probably slightly iterated upon) will then come out as A17.

At least this is the theory that explains why there are so few changes in the A16 we got.

In addition to issues with the move to 3nm, supply chain issues also have to play some role in that, especially given that the iPhone 14 and 14 Plus are using the A15 Bionic from the 13 Pro/13 Pro Max instead of anything badged as an A16.

 

[leman]

In addition to issues with the move to 3nm, supply chain issues also have to play some role in that, especially given that the iPhone 14 and 14 Plus are using the A15 Bionic from the 13 Pro/13 Pro Max instead of anything badged as an A16.

I wouldn't be surprised if the supply chain issues are related to LPDDR5 6400. Apple needs a huge amount of that RAM for their products.

 

[dugbug]

Well supply chain issues probably hit TSMC really hard trying to start a new node at production levels. They needed to build a massive manufacturing pipeline for their biggest customer. The shortage is probably what made them signal the delay to begin with rather than something technical.

 

[sam_dean]

Apple execs on M2 chips, winning gamers and when to buy a Mac | TechCrunch

Apple's M series chips were incredibly well telegraphed when they arrived in late 2020. Apple had been designing its own silicon since the A4 appeared in

techcrunch.com

Other evidence/speculation include:

• Rumors were that M2 SoC was ready much earlier but was held back by the late new Air design.
• Reports and video file name suggest that the M2 Pro/Max were ready as early as October 2022, which would have made it one year since M1 Pro/Max
• M2 used A15 cores (not A16 in iPhone 14), which suggests Apple wants to take advantage of the iPhone SoC design update each year
• Reports are that M2 will be a short-lived generation because M3 is expected to launch on time
• Many speculators point to iPad SoCs, which were not updated yearly. However, the base M chips go into far more devices than old iPad SoCs. This makes it more economical to design and launch new SoCs yearly.
• M1 launched in the same quarter as A14, which suggests Apple develops the base M and the A series in tandem
• Apple has been fighting supply chain issues in the last 3 years, which caused many delays in new designs shipping.

Why is this important? Because gaining 10-20% in performance every two years is not impressive. But once a year? Now that's impressive and exciting.

I hope all your rumors come true. It sounds believable or a very intelligent guess.

If Apple's supply chain people can sync it that way it would be nice.

Predictable replacement cycle of every 11-13 months.

Every quarter has a new product or product refresh so no disappointing quarters for shareholder.

Improving raw performance comes down to a combination of these conditions

- being at the leading process node
- increasing die surface area
- increasing clockspeed
- increasing TDP
- increasing core counts
- increasing transistor counts
- use the latest iPhone chip's cores as base of Mac chips

I'd love to see

- M Ultra chip (2x100W M Max chips) in a MBP with a 240W USB PD charger hopefully at a <$2k premium
- M Extreme chip (4x100W M Max chips) in a Mac Studio or Mac Pro with that consumes ~415W at a <$4k premium

>90% of the R&D money for Apple Silicon chips comes from iPhone profits... <10% of R&D money is from Mac profits.

For battery-powered devices I prefer longer reduced power consumption and improve performance per watt over raw performance.

For outlet-powered devices I prefer raw performance and improved performance per watt over reduced power consumption.

Assuming the binned part can handle the increased power input. I still find it odd that desktops have identical benchmark results as laptops because they use the same exact SoC.

Another product I'd like to see is iPhone chips finding themselves into $699 Macbook 12" and a smaller $299 Mac mini. Both would have 8GB RAM & 256GB SSD. All Mac SKUs would have 2x RAM & 2x SSD.

If it werent for the competition within the smartphone space I doubt Apple Silicon would be that good relative to x86.

The trend seems to be that it is becoming more efficient for smartphone vendors to scale up their products into tablets, laptops & desktops, than it is for desktop vendors to scale down their products into laptops, tablets and smartphones.

Laptops and desktops just have too much headroom when it comes to power consumption.

So when you scale up smartphone tech to the headroom of a PC then you impress even when you still sip power.

 

[Warped9]

I’m sooo eager to see that M3 iMac since it really does look like there won’t be an M2 version.

I’m also interested to see what is next for the Mac Studio and if there will be a new larger iMac.

Finally Apple really should release a smaller and/or less expensive companion to the Studio display.

 

[Onimusha370]

Ever since we had to wait till June 2022 for the M2 (A15X), I've firmly believed that we're on an 18 month cycle for Apple silicon Macs, just as we have been on the iPad Pros for many years. This means that they skip 1 in 3 A series variants essentially, i.e. A14, A15... skip the A16, then A17. If Mac updates became annual then pretty much the whole of Apple's hardware business would be on an annual (predictable) cycle. Don't get me wrong, I'd personally love that, but I think the 18 month cycle feels more natural for how they like to release Macs these days. Fully expecting M3 to launch in October/November 2023 based on the A17 (3nm) and the Pro/Max chips to follow in spring 2024.

 

[leman]

I'd love to see

- M Ultra chip (2x100W M Max chips) in a MBP with a 240W USB PD charger hopefully at a <$2k premium
- M Extreme chip (4x100W M Max chips) in a Mac Studio or Mac Pro with that consumes ~415W at a <$4k premium

I can assure you that you won't see any of these as none of them make sense.

 

[sam_dean]

I’m sooo eager to see that M3 iMac since it really does look like there won’t be an M2 version.

I’m also interested to see what is next for the Mac Studio and if there will be a new larger iMac.

Finally Apple really should release a smaller and/or less expensive companion to the Studio display.

That means a delay of 1-2 years more.

I would not want to wait further for a iMac 27" replacement.

 

[sam_dean]

I can assure you that you won't see any of these as none of them make sense.

Why not?

The 240W USB PD charger is there

The 1.5kW PSU is there.

So long as future SoC can physically fit into the form factors... why not?

There is a market for those SoC performance targets.

 

[James Godfrey]

Personally I expect 18 months cycles with the standard M series chips first, then 9 months later the Pro Max and Ultra iterations then 9 months later back to standard M series and so on…. I fully expect M3 towards the end of the year, with M3 Pro Max and Ultra likely early to mid next year.

 

[sam_dean]

As I’ve said in other posts:

1. I sincerely hope this is the case as I am eager to see the return of annual iMac and MacBook Air updates and biennial Mac Pro and MacBook Pro updates.

2. It’s not just Apple Silicon I am concerned about. Although Apple’s, current designs are excellent, and some of the best they’ve ever had, there is definitely room for improvement, especially in the Mac Pro tower.

3. My fear is that Apple is spending greater RMD effort, both financially, and mentally, on Services, Wearables, and the iPhone. Although those are of great importance to Apple revenue, the Mac is the core business model which holds up the empire. Remember, when the iPhone was first announced, Steve Jobs made a great deal of noting that the iPhone was running a full version of OS X. I recognize that, today, you don’t need the Mac to sell the iPhone. However, the Mac is where Apple’s customer loyalty is most visible. Furthermore, the Mac continues to gain market share, especially among high school and college students. The revenue percentage makes sense as people just don’t update their computers as often as their phone.

4. I’m not suggesting that Apple reduce R&D on the iPhone, I am suggesting they place the same importance on the Mac as they do on the iPhone.

The world is moving at an incredibly fast pace and to see the entire product line of desktops be outdated isn’t going to past muster. It was infuriating with the prior MacPro. Seven years with no update! Apple should’ve delivered a MacPro M1 last year. To spend real money on an M2 when the M3 is around the corner will give many pause. This is what unpredictability does, you don’t know how to plan; and for professionals, using the machine for our income, that is highly upsetting.

1) If OP's points are the actual supply chain target of Apple then all Apple hardware will end up having an annual refresh cycle because the Mac SoC will be refreshed based on the September iPhone chip's cores.

2) I don't think the Mac Pro merits a case redesign. The users would rather prefer Mac chips in them to reduce any futher delays.

3) The Mac business unit contributed the least in terms of revenue at. For 2022 they were estimated to have shipped 27.911 million units worldwide vs 226.4 millon iPhones in the same year.

4) As I see it >90% of R&D chip spend is financed by iPhone sales. <10% of the R&D chip spend is finance by the specific device like the Mac for its specific needs. Like say the 5G modem, WiFi 7, etc may eventually use chips through the effort of Apple R&D.

Mac Pro's turning 4 by December. It may enjoy annual product refresh if Apple's supply chain can sync M3 series to this September 2023's A17 Bionic 3nm chip. Would not be surprice if its raw performance matches those of the November 2020 M1 5nm chip but at a fraction of the power consumption and SoC package size.

Doing this would put iPhone, Mac, iPad, Apple TV, Watch & speakers SoC on the same process node as time moves on within 12 months.

A "refined" 3nm would commence over the following 12 months.

Repeat with 2nm by September 2025 with the M4 series of SoC with Apple buying said fab capacity.

This would help with the economies of scale on a specific process node like 3nm. Apple could book out the whole fab capacity longer so no one else gets the latest node. This is a competitive advantage. Apple could inversely sell their capacity to rivals if they wanted to and if the contract allows for sub-selling.

So say by September 2023 the iPhone chip will have new CPU-core, GPU-core and other transistor tech on 3nm node then Mac, iPad, Apple TV, Watch & speaker SoC will be refreshed based on that iPhone chip as the months goes on over a period of 12 months like clockwork.

Then repeat every 12 months. Predictability will help Apple's supply chain do better.

 

[leman]

Why not?

The 240W USB PD charger is there

The 1.5kW PSU is there.

So long as future SoC can physically fit into the form factors... why not?

There is a market for those SoC performance targets.

Because:

a) Apple is not interested in making a 120W+ laptop (they have had plenty of opportunities to do so in the last 15 years but never took them)

b) Apple won't give you an end-of-the-line professional workstation for $4K

Anyway, let's keep this thread on topic please. There are already other threads where you guys are discussing your unrealistic fantasies and you can always open more of those.

 

[Kazgarth]

Obviously they are gonna update it yearly, all the competition (Intel, AMD and Qualcomm) are doing so.

 

[sam_dean]

The Mac Pro redesign, could have been forgiven for the M1 iteration, which never happened. However, the 2019 Mac Pro tower is a huge monstrosity that could be significantly reduced with the efficiency of Apple Silicon.

They need the space for the PCIe expansion slots and some uses want the physical dimension of the tower.

Similar to why the Mac mini enclosure did not reduce in volume to fit the iPhone-sized Mac mini logic board.

 

[Jack Neill]

I can still use 2012/2013 Macs in my daily flow and I hope that my M1/16/1Tb Air can last for 10 years and be just as useful. I would have gotten my moneys worth of 944$ I paid for it. By then I'll be ready for an M10 Pro.

 

[sam_dean]

Because:

a) Apple is not interested in making a 120W+ laptop (they have had plenty of opportunities to do so in the last 15 years but never took them)

b) Apple won't give you an end-of-the-line professional workstation for $4K

Anyway, let's keep this thread on topic please. There are already other threads where you guys are discussing your unrealistic fantasies and you can always open more of those.

Apple does make 140W laptops... the MBP 16" M1 Max.

Apple has the Ultra that fits the 240W USB PD charger.

Only reason it was not done with the 2021 MBP 16" is possibly the SoC package's size, charger spec wasn't published yet or not enough supply of the SoC.

 

[Confused-User]

I never said that A16 was back ported from 3nm. What I am saying that I suspect that A16 is not actually A16. I almost believe that original A16 was supposed to be a 3nm chip with all new internal design but since 3nm was delayed Apple has scrambled toughener another variant of A15 and called it "A16". The actual A16 (probably slightly iterated upon) will then come out as A17.

At least this is the theory that explains why there are so few changes in the A16 we got.

This is pretty much a summary of stuff I (and others, probably including leman) have previously posted. There is a great deal of evidence for this. Whether it's actually true or not, we may never know. Though if we do get an M3 Mac Pro in the next 5 months (which I suspect will happen) it will make this almost certain. Another thing that would make this a near certainty is if we get an A17 with a really big performance boost beyond just the advantage given by the move to N3B, showing two years of development (the original A16 design, plus further progress over the last year).

 

[sam_dean]

I can still use 2012/2013 Macs in my daily flow and I hope that my M1/16/1Tb Air can last for 10 years and be just as useful. I would have gotten my moneys worth of 944$ I paid for it. By then I'll be ready for an M10 Pro.

2012/2013 Mac on a 22nm process node to a 2020 Mac on a 5nm process node!

Upgrade after last Security Update and for the purpose of preventive maintenance!

After each decade-long use you'd really feel the difference.

What we want is Apple always perpetually ahead in performance per watt.

 

[aeronatis]

No, 3nm was originally planned for 2022 and was delayed by TSMC. I’m fairly sure that A16 and M3 were supposed to be 3nm designs with completely new CPU and GPU (we even have parent details on the GPU).

In the end, what Apple shipped as A16 is likely a placeholder, basically an A15 that has been slightly tweaked for higher clocks.

I agree. This is in line with the rumours of that iPhone 14 Pro SoC was originally intended to have RT cores and they had to cancel it afterwards as the power consumption was too high with 5 nm design.

Apple does make 140W laptops... the MBP 16" M1 Max.

Apple has the Ultra that fits the 240W USB PD charger.

Only reason it was not done with the 2021 MBP 16" is possibly the SoC package's size, charger spec wasn't published yet or not enough supply of the SoC.

Apple doesn not make 140-watt laptops. 140 Watt charger is only there to charge the laptop faster. MacBook Pro 16" M1 Max consumes less than 100 watts at full CPU + GPU load. That is how they keep the performance same both plugged in and on battery. I believe Apple would like to keep that as is.

I agree M Ultra chip would fit inside a laptop chassis as there are tons of laptops consuming even more than that. However, that would take away all the advantages of Apple Silicon laptops like quiet operation, max performance on battery, minimal performance loss at sustained load, etc.