M2 Ultra in a Macbook Pro will be possible at 3nm or 2nm

[sam_dean]

The known technical limit would be 240W USB PD charger.

Top-end Mac Studio M1 Ultra 5nm has a max power input of 215W without a display.

A M2 Ultra 3nm with at least a 20% performance per watt improvement would allow an Ultra in a MacBook Pro with 240W charger.

The question would be how big of a market is there for laptop that sells for nearly $7,000 and weighs nearly 2x 4.7 pounds (2.1 kg).

Mac Studio M1 Ultra's heat sink fan was made of a more thermally conductive material that weighs more so it can maintain its heat sink fan's RPM and physical form factor dimension.

As the M1 Ultra chip is 2x the die/surface area of a M1 Max chip then the heat it generates would be 2x as well. So that Macbook Pro M2 Ultra 3nm would run hotter than the top-end Macbook Pro 16 M1 Max.

 

[Xiao_Xi]

Wouldn't such a laptop go against the image Apple wants to sell of its laptops: light, cool and quiet?

 

[Kazgarth]

The M2 is not 3nm, it's N5P. And it will have more core count.

M2 Max has 12 cores from the leaked GB 5 (vs 10 in M1 Max), so the Ultra will have 24 cores (2x) if they plan to do M2 refresh.

The gradual core count increase with every M generation is enough to maintain the MBP performance crown, without unrealistic sacrifice to battery life as OP suggested.

 

[sam_dean]

Wouldn't such a laptop go against the image Apple wants to sell of its laptops: light, cool and quiet?

2019 MBP 16" vs 2021 MBP 16" somewhat reversed this trend

 

[Lihp8270]

I can’t see the M2 max / ultra being on a different node size to the M2.

After all, the M1 Max, pro, ultra etc were “stitched together” multiples of an M1.

Having the M2 pro/max/ultra be different to M2 doesn’t make sense from the naming.

I’d expect them to be using the same node as the M2 or have a different name

 

[leman]

I very much doubt that Apple will ever release a laptop with higher thermal design power than 80-90 watts. They have long determined that this is the sweet spot for the 15-17" form factor and never deviated from this principle in over a decade.

USB DP is just for fast charging. Apple isn't interested in making 100W+ laptops.

 

[maflynn]

A M2 Ultra 3nm with at least a 20% performance per watt improvement would allow an Ultra in a MacBook Pro with 240W charger.

Except that the M2 is not built on 3nm. The Pro/Max/Ultra are just bigger variants of the M1, so its safe to assume that will be the case for the M2.

 

[TSE]

Except that the M2 is not built on 3nm. The Pro/Max/Ultra are just bigger variants of the M1, so its safe to assume that will be the case for the M2.

I agree. All the tech youtubers that are claiming that the next M2 Pro/Ultra/Max are all 3nm are goofballs.

The M2 Pro/Max/Ultra are going to be enhanced 5nm like the M2/A15, which means 10-15% faster CPU, 20-35% faster GPU.

The M3 and it's variants are going to be based on the A16 - so 4nm which should see about 20% better efficiency.

 

[maflynn]

The M2 Pro/Max/Ultra are going to be enhanced 5nm like the M2/A15, which means 10-15% faster CPU, 20-35% faster GPU.

Yep, and the leaked Geekbench scores are supporting this. The M2 was always just a spec bump of the M1, and so we'll not see anything on 3mn this year or next.

'M2 Max' Geekbench Scores Leak Online, Revealing Rumored Specs and Performance

 

[Kazgarth]

I agree. All the tech youtubers that are claiming that the next M2 Pro/Ultra/Max are all 3nm are goofballs.

The M2 Pro/Max/Ultra are going to be enhanced 5nm like the M2/A15, which means 10-15% faster CPU, 20-35% faster GPU.

The M3 and it's variants are going to be based on the A16 - so 4nm which should see about 20% better efficiency.

Exactly, plus they don't realize how expensive is N3 silicon is from TSMC.
It's not gonna be used for something larger than an iPhone chip for a long time (and mostly likely only the iPhone Pro margins can offset the insane cost of N3 silicon for the first 1-2 years).

 

[leman]

Exactly, plus they don't realize how expensive is N3 silicon is from TSMC.
It's not gonna be used for something bigger than an iPhone chip for a long time (and mostly likely only the iPhone Pro margins can offset the insane cost of N3 silicon for the first 1-2 years).

View attachment 2124300

Are the increased wafer costs really a problem for high-margin products like the MacBook Pro? We are talking about $30-50 price increase for the largest chips. Doubt it will be an issue.

 

[Kazgarth]

Are the increased wafer costs really a problem for high-margin products like the MacBook Pro? We are talking about $30-50 price increase for the largest chips. Doubt it will be an issue.

It's not that simple.
The larger the chip is the more expensive the yield loss (% of defective chips) per wafer is. On top of higher cost of the new node, it compounds to way, way greater amount than "$30-50".

 

[saintmac]

2019 MBP 16" vs 2021 MBP 16" somewhat reversed this trend

not so much. It is slightly heavier but completely cool and quiet in most scenarios where the 2019 wasn't.

 

[leman]

It's not that simple.
The larger the chip is the more expensive the yield loss (% of defective chips) per wafer is. On top of higher cost of the new node, it compounds to way, way greater amount than "$30-50".

View attachment 2124313

Assuming the yield and chip size are comparable (which is a problematic assumption, I know), you will roughly get the same amount of chips per wafer. So the expected price increase is around 25%, same as the wafer cost. I doubt that a M1 Max die costs hundreds of dollars to manufacture. Let's take the image you post as a rough baseline. An M1 Max is a 20x20mm chip, assuming 136 working dies per wafer that's $136 on N5 and $147 on N3. This is pretty much negligible compared to the cost of the package assembly and RAM. And of course, we are talking about machines that start at over $2000.

 

[Pressure]

Exactly, plus they don't realize how expensive is N3 silicon is from TSMC.
It's not gonna be used for something larger than an iPhone chip for a long time (and mostly likely only the iPhone Pro margins can offset the insane cost of N3 silicon for the first 1-2 years).

View attachment 2124300

I don't think that's true though.

The increase is even less than going from 7nm to 5nm. The wafer prices are already "priced in" with regards to Apple's internal road map and is probably less than $20 per chip for the larger ones, like the M1 Max or machines that starts at $2000 or more.

 

[jdb8167]

Yep, and the leaked Geekbench scores are supporting this. The M2 was always just a spec bump of the M1, and so we'll not see anything on 3mn this year or next.

'M2 Max' Geekbench Scores Leak Online, Revealing Rumored Specs and Performance

That doesn't make any sense unless you think TSMC isn't really ready with N3 yet despite their press releases saying that they are going into production 4th quarter this year. Why would it take more than another year for Apple to release a Mx SoC using N3?

TSMC FINFLEX™, N2 Process Innovations Debut at 2022 North American Technology Symposium

 

[Kazgarth]

Assuming the yield and chip size are comparable (which is a problematic assumption, I know), you will roughly get the same amount of chips per wafer. So the expected price increase is around 25%, same as the wafer cost. I doubt that a M1 Max die costs hundreds of dollars to manufacture. Let's take the image you post as a rough baseline. An M1 Max is a 20x20mm chip, assuming 136 working dies per wafer that's $136 on N5 and $147 on N3. This is pretty much negligible compared to the cost of the package assembly and RAM. And of course, we are talking about machines that start at over $2000.

I would say it's safe to assume that the yield of a brand new node is lower than a 3 years old N5. Again your calucluation assumes zero yeild differences between the old and the new node. Plus consider the fact in my next reply.

I don't think that's true though.

The increase is even less than going from 7nm to 5nm. The wafer prices are already "priced in" with regards to Apple's internal road map and is probably less than $20 per chip for the larger ones, like the M1 Max or machines that starts at $2000 or more.

The graph I've posted reflects the wafer prices as of November 2022 and that includes the recent %10-20% price hike to both 7nm & 5nm from TSMC.

Tech Wire Asia

Tech Wire Asia is an independent tech news website that covers enterprise and technology in Asia

techwireasia.com

But as we know Apple was the first to secure TSMC's N5 contract back in 2020 before the recent price hikes for new customers, so the actual price jump for them is more than what the graph imply.

 

[darngooddesign]

The question would be how big of a market is there for laptop that sells for nearly $7,000 and weighs nearly 2x 4.7 pounds (2.1 kg).

You forgot to add always on fans, poor battery life due to airline limitations restricting its capacity, and the probability that it would likely still throttle while trying to use the Ultra's performance unless the fans were roaring. Most thick and heavy PC laptops are gaming machines, but a MacBook doesn't even have that benefit. An Ultra MBP would probably feel like using a hot and loud Intel 16" MBP.

Even though it would likely have a larger 18" screen, probably not much of a market.

Then again, these are probably acceptable tradeoffs for the right person.

 

[Wokis]

A Macbook Ultra.. I like it! Very niché, but a lot of production people would probably say it changed their lives.

17", or even 18". A bit chonkier. Whatever, the right tool for the job etc.

 

[JPack]

Exactly, plus they don't realize how expensive is N3 silicon is from TSMC.
It's not gonna be used for something larger than an iPhone chip for a long time (and mostly likely only the iPhone Pro margins can offset the insane cost of N3 silicon for the first 1-2 years).

View attachment 2124300

And wafer prices are only part of the cost. The most expensive stuff is EDA, verification, and mask sets. All that costs nearly a billion dollars.

How many units of MBP M2 Ultra would Apple need to sell to recoup a billion dollars? Apple can't just take an M2 Ultra chip for the desktop and put it in a notebook. They would tune it.

Apple tosses in A14/A15 in everything from iPhone to iPad to Apple TV because it helps amortize R&D costs. You can't throw in M2 Ultra in iPhone SE.

 

[deconstruct60]

The known technical limit would be 240W USB PD charger.

Top-end Mac Studio M1 Ultra 5nm has a max power input of 215W without a display.

A M2 Ultra 3nm with at least a 20% performance per watt improvement would allow an Ultra in a MacBook Pro with 240W charger.

Flapping your arms and hand waving at "3nm" magic won't change the problem that the package won't fit in a current MacBook Pro 16" chassis (let alone the 14" one).

MacBook Pro 16" 2021 Logic Board Replacement

Use this guide to replace the logic board in your MacBook Pro 16" 2021. Note that Touch ID will not function after replacing the logic board. The...

www.ifixit.com

Even is used extra TSMC N3 magic pixie dust to shrink the Ultra down to the same size as a Max package, the Ultra requires more LPDDR5 RAM packages to 'surround' the magic N3 package to feed it the required bandwidth of data. The Memory isn't going to shrink with N3. N3 (and later) will shrink the compute logic substantially but analog (and external pins to relatively distance other packages and I/O) and memory ( RAM , cache ) aren't going to shrink as well. ( there is 3D stacking of RAM dies to drive more capacity in a same footprint , but the ram package footprints are off smaller footprint trendline now. )


Trying to stuff an Ultra die where the associated RAM can't go is relatively pointless. The Max and required RAM are actually physically hemmed in by the fans. So even if keep the power dissipation constant, there really isn't a 'win'. ( in space saved or heat/power (more RAM will be more power) )

The question would be how big of a market is there for laptop that sells for nearly $7,000 and weighs nearly 2x 4.7 pounds (2.1 kg).

If you about to make another "laptop" chassis that is 'luggable' then why bother. The Ultra version of the Studio is just 8 lbs. It would fit in an ordinary large suitcase (or similarly sized more custom travel container) . It isn't a huge piece of equipment to transport from one desktop to another.

The "Max" version is 5.9 pounds. If sprinkle N3 magic dust and drop the power down closer to the max range could get away with the aluminum cooler for both. And even closer still to the 4.7 pounds that is supposedly tolerable to travel with.

 

[Tyler O'Bannon]

M2 series will likely all be the same node. RAM seems to be 1.5x’ing, and CPU and GPU core count increasing.

There will be an M3 or M4 Max that probably beats the M1 Ultra, but there probably won’t be an Ultra chip in MBP. The heat sink for ultra is too much.

 

[Alex Cai]

A Macbook Ultra.. I like it! Very niché, but a lot of production people would probably say it changed their lives.

17", or even 18". A bit chonkier. Whatever, the right tool for the job etc.

Then why don’t they carry a Mac Studio and a portable battery?

 

[maflynn]

A Macbook Ultra.. I like it! Very niché, but a lot of production people would probably say it changed their lives.

You think? I have no way to know either way, but I suspect between the cost, the size, weight and likliehood of throttling - the appeal would be miniscule.

The R&D, and production costs would be hard to justify it when the target audience is tiny.

Also the battery life as others mentioned would be horribly short and battery life is one of the major tent poles Apple markets their laptops. I mean, right now, I can run my 14" MBP on battery and its performance will be that as if it was plugged in. I'm not sure that would be possible with an ulta, and if it was, the battery life would be horribly short.

All speculation to be sure, but I just don't see much of a market for a M2 Ulta MBP.

 

[darngooddesign]

Then why don’t they carry a Mac Studio and a portable battery?

Carrying all that plus a mouse, keyboard, and display is significantly more cumbersome than even a 7#, 18" laptop. You own a MBA instead of a MacMini for the same convenience reasons.

There's probably a small, niche market for such a MBP-Ultra. If we weren't in a component shortage Apple might choose to serve it, but right now they won't.