Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Two scenarios other than the Mac Pro being discontinued…

1) The m3 ultra may be reserved for the Mac Pro moving forward. The studio will get the M3 pro and M3 max moving forward.

This basically ignores the Mini Pro. Mini with M3 Pro and Studio with M3 Pro ... where is the differentiation?
Apple isn't going to saddle the Studio with a Mx Pro. The entire upgrade in the cooling of the Studio chassis was to make the Mx Max variable in the Studio. Apple isn't going to toss that away at all. Apple already had a Mx Pro viable chassis!!!

Additionally, the M3 Pro's 'extra helping' of E cores versus the Max distinctly points to enclosures with less thermal headroom; not more.

In fact, the Studio would be better if the I/O ports were consistent. 6 Thunderbolt ports no matter which SoC.
Don't need MORE four Thunderbolt port only options! They should have less. Ports where you can easily get to them o the desktop; imagine that *cough* .


2) The studio will get the M3 max, but Apple has the M3 Extreme under development for the Mac Pro and has done a pretty good job of keeping it under wrap.

Better chance that the Mac Pro just skips the M3 all together than do that. The 'quad thing' is broke economically. Without an actual chiplet design it isn't likely going to be viable ( pounding 'round peg' monolithic dies into a big chiplet ('square hole' ) isn't going to reasonably affordable. Nevermind there is really no TSMC CoWoS capacity to make them for a couple years either. Given Apple's MP track record over the last decade a new MP every 12-14 months is plain unlikely. 4-6 years is more they established track record. If Apple got to an every 2-2.5 year cadence they'd be going twice as fast. Pretty good chance Apple isn't a huge hurry to kill the M2 Max die ahead of schedule ( or relatively quickly) if they can get a predictably steady drip flow of MP production that the factories can keep up with at a reasonable cost.

Smaller chance Apple comes with a M3 Ultra and just focuses on their dramatically upgraded GPU implementation to say it is a more solid step up from the MP 2019 and those stock Apple GPU options.
If Apple fixes the PCI-e v4.0 backhaul ( i.e., two x16 feeds) then it will have a bigger gap on the Studio's Thunderbolt limitations. Throw in the other M3 improvements on CPU/GPU/NPU and it is even better than MP 2019 configurations on out of the box performance. ( Apple is going to be in the 'quad M1 Max' range on lots of GPU metrics. )
[ It isn't about being a 4090/7900-killer GPU. It is far more about getting developers to tune up for the new GPU arch optimizations in the Apple software ecosystem, than 'killing' the alternative GPUs. Do not need an 'Extrame' to motivate folks to write more optimized GPU code. ]


Other option for Apple would be to introduce a "Mac add-in Card". No new Mac Pro but being able to put a new Mac inside a enclosure that as an 6 or 8-pin powered PC card. Run a virtual ethernet connection to the host system and folks can just 'remote screen' onto the card from the host system (and SMB file share). Apple needs 'better' value add PCI-e cards to put into a Mac Pro more so than a new CPU. Don't really need a 'new' SoC to do that. They could use the M3 Pro. ( might be even be able to squeeze that into a 75W bus power limit and don't need any 'evil' external power wires that Apple dislikes so much. )
 
Last edited:
  • Like
Reactions: Chuckeee
Does it even make any sense to continue producing a M2 Ultra chip when a M3 Ultra is out?

It could. Depend upon who is the bigger consumer of Ultra chips. if the Ultra Studios sell 5:1 ratio to Mac Pros then the M2 Max's could be mostly 'paid for' dies (for the overhead and ROI. ) . Apple stockpiles a relatively smallish amount and dribbles them out for 5-10 months. As long as the Mac Pro run rate is steady and large nough to support a contractor production ... it is just fine. (e.g., Mac Pro 2013 ran for years on a steady flow of rackmount and 'just large enough' other niche interest folks to keep the line going. )

Reversed that doesn't make sense. Studio Ultras 1:5 and only making very few M3 Ultras ... that would be the one that makes no sense.

You are trying to kill of the M2 Max faster!! Why would Apple want to kill a relatively large die faster to jump to one that is far more expensive to make. Pretty good chance Apple would like to eek out more margin on a mature Soc if they can. The bulk of the product line up is built that way. They sell same SoC for many multiple years. They don't drop them as fast as possible in general.

The substantive problem with the Max and Ultra is that there is no "hand me down" other Mac Product to put them into. Relatively rapidly kill off your largest dies is a pretty good way to making less money , not more.
( look over in large die GPU market. Few are trying to kill those off on strict 12 month cycles at all. )


Pretty good chance part of the Mac Pro's job is to soak up larger dies for longer periods of time. A product that had a 'normal' 4-6 year cadence could/would do that. If Apple shrank that down to 2-2.5 years it still would do that.
 
Don't forget that Apple kept selling the old trashcan Mac Pro for almost 6 years without upgrading it.

I expect that they will upgrade the current Mac Pro occasionally or maybe never, but will continue to sell it. They can drop the price over time, or not, depending on how locked-in the businesses are that need it.

On general practice, Apple doesn't do price drops over time.
The MP 2013 only got a price drop when they started to get embarrassed in the press when it was close to to 1000 day mark with no updates. By 2017-19 , the last two years of that term, for many the iMac Pro was the replacement for normal desktop users ( and MP got a discount to limp along for an extra two years). 2017 was iMac Pro introduction, so of that 2017 MP drop was to clear some pricing headroom for iMac Pro.

MP 2019 when a good 4 years with no formal price drop at all. MP 2009-2012 no substantive drop either.

Highly doubtful Apple would go that route. (as Apple had to chop off the end of vintage/obsolete support time after sales stopped for the MP 2013 abnormally quickly. Doubtful they'll want to make that a reoccurring practice). Also a very different era when AMD was almost comatose in workstation CPU updates and Intel was milking the segment. Not happening now.

Could there be some defacto constant sales ( e.g., iPad Gen 9 on sale at Amazon and a few others for $249=-269 for about 1/2 the year. )? But Apple, probably not.

Apple will do product price adjustment if the current exchange rate goes bad against the USA dollar, but in constant dollars it tends not to move throughout the life of Mac product. ( not really the iPhone model where sell 2-3 years old stuff side by side with new stuff. Again , another model is pushing the price cut... not just sell an 'older' thing. )
 
Apple should have kept the Mac Pro on the latest Intel hardware for a few more years and then quietly discontinued it.

Not viable. Each year Apple is dropping 10's of millions of Intel Macs onto the Vintage/Obsolete list. The number of macs that macOS Intel supports is getting steadily smaller. The Mac Pro just by itself is not a viable software ecosystem. A substantial number of 3rd party software vendors are going to drop out once get to critically small remnant left.

macOS on Intel very likely isn't dying this year or next , but it is in very substantive decline. Apple doesn't have 'years' to keep selling a macOS Intel following by their standard "5-7 years after stop selling". Apple is chopping off the Intel models at a pretty strict 5 years. Already down to 4.5 at best now (last Intel model dropped mark).

The number of "Mac Pro only" software titles is very, very , very , very small. That's the size of ecosystem you are pointing at. Once the other Mac are gone what is left is not viable. The best selling macs have long been completely off Intel. Those were the real primary support to make macOS Intel long term viable.
 
I can't be the only person who needs a Mac with Nvidia 4090 level GPU performance…? I don't really care what form factor it's in, I just need a beefy workstation for Cinema 4D and Adobe software, and I really don't like using Windows…

I'm pushing on with an M2 Max for 2D work and a 2019 Mac Pro full of GPU's for 3D work… but not ideal.
 
The M3 Ultra Mac Studio will be my next Mac, replacing a 2019 i9 MacBook Pro for daily use (with 64GB RAM and a 2TB SSD installed, the MBP is still very, very powerful and a really nice mobile machine).
The M3 Ultra Max Mac Pro Extreme is gonna be my next purchase. It's rated at many several flops, bits, and bytes. Hertz? It's got...all of them. Or enough?
 
My probably meaningless response to this.
Mac Pro is limited. Apple needs to pull its head out of its.....you know.
Nice case though.
 


A new Mac Studio model with the M3 Ultra chip—which could be more powerful than expected—will launch in mid-2024, according to a new report.

M3-Mac-Pro-and-Studio-Feature.jpg

The report comes from Taiwanese research firm TrendForce citing ICsmart. The current Mac Studio, which contains M2 Max and M2 Ultra chip options, was introduced at WWDC in June 2023. With TrendForce forecasting another mid-year update for the Mac Studio, a repeat appearance at WWDC seems likely this year.

The Mac Pro is the only other Mac model that contains an "Ultra" Apple silicon chip, so its absence from today's report discussing the M3 Ultra is noticeable. While, logically, the Mac Pro should receive an update to add the M3 Ultra chip alongside the Mac Studio, just like it received the M2 Ultra chip last year, there have been no rumors to suggest this will be the case again this year as of yet.

Interestingly, TrendForce claims that the M3 Ultra chip will be fabricated with TSMC's N3E node, just like the A18 chip that is expected to debut in the iPhone 16 lineup later in the year. This means it would be Apple's first N3E chip. The M3, M3 Pro, and M3 Max chips are fabricated with TSMC's N3B process, just like the A17 Pro. N3E is an enhanced version of TSMC's 3nm process, offering slightly better performance and higher production yield.

It would be somewhat odd for Apple to make the M3 Ultra its first N3E chip, since Apple's "Ultra" chips comprise two "Max" chips linked via its UltraFusion chip interconnect technology, meaning that an N3E M3 Ultra would effectively be two M4 Max chips. To date, each generation of Apple's custom silicon chips have used the same fabrication process across the lineup, since the chip architecture is simply scaled up. Using different nodes in the same family of chips, especially on the "Ultra" chip variant, would be an unprecedented move.

The only explanation if Apple does opt for this route would be that the company plans to introduce the M4, M4 Pro, and M4 Max chips sooner rather than later, perhaps in late 2024, and would want to ensure that the M3 Ultra Mac Studio remains competitive with the newer chips. Even so, in this instance, it would surely be more straightforward to brand the chip "M4 Ultra" and skip the M3 Ultra entirely.

As a result, this part of the report should be treated with some skepticism. Nevertheless, a mid-year time frame for the launch of a new Mac Studio seems likely.


Article Link: M3 Ultra Mac Studio Rumored to Launch in Mid-2024, But No Sign of New Mac Pro
I'm hyped for the new Mac Pro which combines the best of the original 2011 model with the current model.

MacPro.jpg
 
I can't be the only person who needs a Mac with Nvidia 4090 level GPU performance…? I don't really care what form factor it's in, I just need a beefy workstation for Cinema 4D and Adobe software, and I really don't like using Windows…

I'm pushing on with an M2 Max for 2D work and a 2019 Mac Pro full of GPU's for 3D work… but not ideal.
You probably aren't, but the question is whether you remain a market large enough for Apple to consider worth serving.

I am starting to realise that the main advantage of Apple Silicon isn't so much performance but power efficiency. This makes them a natural fit for laptops, but much of their strength evaporates in a desktop where users have differing priorities. The problem then comes when the Mac Pro is already a very small demographic, and I don't see Apple going out of their way to come up with a custom chip just for it that prioritises raw performance. The sheer amount of R&D involved is just not worth it.

And since Apple sells way more laptops than desktops, this may be a sacrifice they are willing to make.
 
  • Love
  • Like
Reactions: smulji and Chuckeee
This was supposed to have been based on sales September 2022 quarter (M1 generation).
As I've pointed out before about those numbers, the calendar period it covers is when the Mac Studio was new. It is highly likely that the last quarter of 2023 will show significantly different distribution of models.

If YouTube videos are anything to go by (and they very well could not be), then the Mac Studio has gotten a lot of attention. I even wonder if Apple sells more of them than the Mac Mini.
 
  • Like
Reactions: Chuckeee
Either way, the Mac Pro is dead.
I'm thinking the Day of the Desktop is over.

For now Apple will push the Mac Studio (and matching Studio Display) as the choice for software developers, to develop applications for Apple's primary business - portables/wearables.

That is why the iMac got the lightest-touch-ever makeover (a really simple board change), and why the Mini is so oddly spec'd (the base model is less than half the cost of the 'Pro SOC model). The 2023 Mini offerings struck me as a bit of a departure for Apple, offering the M2 Pro and a different HDMI version, but only giving the 'Pro one memory upgrade path. The M2 Pro Mac Mini isn't even less expensive than the Studio, once one adds in a memory upgrade, the 10GBe upgrade, and buy an external USB hub with an SD card slot.

It all suggests to me that Apple doesn't really have a long term plan for the desktop market, other than the Studio. The Mac Pro could disappear and the stockholders won't notice any different dividend. Apple could just drop the Mini too, and only offer an iMac and the Studio... until the iMac is replaced with a larger iPad Pro on a stand, leaving only the Studio.
 
Last edited:
  • Love
  • Like
Reactions: smulji and Chuckeee
It could. Depend upon who is the bigger consumer of Ultra chips. if the Ultra Studios sell 5:1 ratio to Mac Pros then the M2 Max's could be mostly 'paid for' dies (for the overhead and ROI. ) . Apple stockpiles a relatively smallish amount and dribbles them out for 5-10 months. As long as the Mac Pro run rate is steady and large nough to support a contractor production ... it is just fine. (e.g., Mac Pro 2013 ran for years on a steady flow of rackmount and 'just large enough' other niche interest folks to keep the line going. )

Reversed that doesn't make sense. Studio Ultras 1:5 and only making very few M3 Ultras ... that would be the one that makes no sense.

You are trying to kill of the M2 Max faster!! Why would Apple want to kill a relatively large die faster to jump to one that is far more expensive to make. Pretty good chance Apple would like to eek out more margin on a mature Soc if they can. The bulk of the product line up is built that way. They sell same SoC for many multiple years. They don't drop them as fast as possible in general.

The substantive problem with the Max and Ultra is that there is no "hand me down" other Mac Product to put them into. Relatively rapidly kill off your largest dies is a pretty good way to making less money , not more.
( look over in large die GPU market. Few are trying to kill those off on strict 12 month cycles at all. )


Pretty good chance part of the Mac Pro's job is to soak up larger dies for longer periods of time. A product that had a 'normal' 4-6 year cadence could/would do that. If Apple shrank that down to 2-2.5 years it still would do that.
Are you aware of how Apple handle their Mac chip productions? It isn't the same as Intel or AMD.

Ultra chips are derived from two Max chips. Are M1 Max chips still in production for the past half year?

Is Apple still making M1 Ultra chips today when the M2 Ultra has been out for 7 months already?

It is likely that M3 Ultra chip will be out by WWDC 2024 for both the Mac Studio & Mac Pro.

By then the M2 Max chips that are the basis of the M2 Ultra chip would be out of production.

This is why that rumor of the M3 Ultra chip being Mac Studio-only is bizarre unless Apple has a stockpile of M2 Ultra chips.
 
I can't be the only person who needs a Mac with Nvidia 4090 level GPU performance…? I don't really care what form factor it's in, I just need a beefy workstation for Cinema 4D and Adobe software, and I really don't like using Windows…

I'm pushing on with an M2 Max for 2D work and a 2019 Mac Pro full of GPU's for 3D work… but not ideal.
Question is, are you and the people with your use case numerous enough for Apple to invest in creating a compuer tailored for your needs?
 
Was it actually impossible for Apple to implement expandable memory alongside the SoC? I ask because I assume that the 192GB of unified memory simply won't be enough for those that spec'd their intel Mac Pro with 1.5TB.

So much about the Mac Pro transition to AS feels like a downgrade.
 
Apple Silicon itself is a failure for Mac Pro especially for SoC design. Clearly, Apple doesn't know how to make a workstation just like Mac Pro 2013.
It's not that the Mac Pro is a failure. It's more that the Mac Studio is too good and the needs of a large tower just are not the same anymore. Many of us have long moved on to external devices for expansion since USB3 can handle so much now.

There was a time when expansion cards were the only viable way to expand the features of a computer. That is no longer true today and it's actually rare to need an expansion card over what USB3 and Thunderbolt can provide. Even large raids today used in video production are typically external vs internal. Internal storage just seems rather silly for most users now.

It's a very niche market now that truly needs a large tower to add expansion cards. Again it's not that the Mac Pro is bad. It's that it's not really needed anymore and the Mac Studio takes away a large chunk of the market that once bought the Mac Pro. In the past the only desktop option because an iMac was the Mac Pro. Now there is an equally performing more cost effective option and yeah most users are going to go that route.

If the Mac Pro magically had a M3 Ultra Pro or whatever a next level would be I still don't think it would sell very well. Even the Max level chips provide insane performance for a lot of content creators today and the Ultra already provides a ton of performance for a ton of tasks. I think anything higher than the Ultra would be even more niche now. Apple would invest a ton in a Mac Pro with a very specific one use chip that just would not sell very much at all. That seems like a huge waste to me if 1% of 1% of users would ever actually use it.
 
It's not that the Mac Pro is a failure. It's more that the Mac Studio is too good and the needs of a large tower just are not the same anymore. Many of us have long moved on to external devices for expansion since USB3 can handle so much now.

There was a time when expansion cards were the only viable way to expand the features of a computer. That is no longer true today and it's actually rare to need an expansion card over what USB3 and Thunderbolt can provide. Even large raids today used in video production are typically external vs internal. Internal storage just seems rather silly for most users now.

It's a very niche market now that truly needs a large tower to add expansion cards. Again it's not that the Mac Pro is bad. It's that it's not really needed anymore and the Mac Studio takes away a large chunk of the market that once bought the Mac Pro. In the past the only desktop option because an iMac was the Mac Pro. Now there is an equally performing more cost effective option and yeah most users are going to go that route.

If the Mac Pro magically had a M3 Ultra Pro or whatever a next level would be I still don't think it would sell very well. Even the Max level chips provide insane performance for a lot of content creators today and the Ultra already provides a ton of performance for a ton of tasks. I think anything higher than the Ultra would be even more niche now. Apple would invest a ton in a Mac Pro with a very specific one use chip that just would not sell very much at all. That seems like a huge waste to me if 1% of 1% of users would ever actually use it.
Still a failure when Mac Pro does not have proper hardware.
 
If the Mac Pro magically had a M3 Ultra Pro or whatever a next level would be I still don't think it would sell very well.

Probably not, no.

But, suppose they split the difference: right now, going from the M2 Ultra Mac Studio to the M2 Ultra Mac Pro is $3,000. What if, instead, you could

a) get an M2 Ultra Mac Pro for $2,000 more, or
b) get an M2 Extreme Mac Pro for $4,000 more. Now we'd be talking a $7,999 starting price for the M2 Extreme.

You'd have 48 CPU cores, but perhaps more importantly, you'd have a RAM ceiling of 384 GiB.
 
I checked the original source ( https://www.icsmart.cn/72181/ ) and ran it through Google Translate.

From this:

其中,高通预计将在即将推出的Snapdragon 8 Gen 4 SoC中采用N3E制程代工,而联发科也计划在其下一代旗舰芯片Dimensity 9400上采用N3E制程代工;AMD备受期待的Zen 5 CPU、RDNA 4 GPU和Nvidia的Blackwell服务器GPU,也有望采用台积电N3E制程;另外苹果也将继续在后续的M3 Ultra芯片和iPhone 16 Pro的A18 Pro SoC中使用台积电的N3E制程节点。

苹果M3 Ultra芯片可能会在年中改进后的Mac Studio中首次亮相,


it gave me this (emphasis mine):

Among them, Qualcomm is expected to use the N3E process foundry in the upcoming Snapdragon 8 Gen 4 SoC, while MediaTek also plans to use the N3E process foundry in its next-generation flagship chip Dimensity 9400; AMD's highly anticipated Zen 5 CPU, RDNA 4 GPU and Nvidia's Blackwell server GPU are also expected to use TSMC's N3E process; in addition, Apple will continue to use TSMC's N3E process node in the subsequent M3 Ultra chip and iPhone 16 Pro's A18 Pro SoC.

Apple M3 Ultra chip may debut in revamped Mac Studio mid-year,


Are there arny native speakers on this thread who could check the translation? It mightt just be a bad translation but, if it's correct, it indicates that the original source is itself confused (and thus can't be trusted), since they are saying "Apple will continue to use TSMC's N3E" instead of "Apple will switch to using TSMC's N3E". I.e., the translation indicates they mistakenly think Apple is using N3E now.
 
  • Like
Reactions: DaniTheFox
You probably aren't, but the question is whether you remain a market large enough for Apple to consider worth serving.

I am starting to realise that the main advantage of Apple Silicon isn't so much performance but power efficiency. This makes them a natural fit for laptops, but much of their strength evaporates in a desktop where users have differing priorities. The problem then comes when the Mac Pro is already a very small demographic, and I don't see Apple going out of their way to come up with a custom chip just for it that prioritises raw performance. The sheer amount of R&D involved is just not worth it.

And since Apple sells way more laptops than desktops, this may be a sacrifice they are willing to make.
I know you were responding specifically to a poster asking about the GPU, and there I agree with you

However, separately, I think it's worth noting that it's a different story for CPU.

Apple's max CPU SC and MC performance roughly equals that of the fastest competing desktop processors, and significantly exceeds that of the fastest competing mobile processors. So Apple's CPU stage allows its laptops to significantly outperform Intel/AMD devices, and its desktops to match theirm.

That's of value even to performance-focused desktop user (at least in the CPU space), since it enables the same CPU performance in a much quieter and cooler box.

At the same time, I think Apple could do more with both its CPU and GPU desktops, but allowing them somewhat higher clocks than they use in their laptops (something they've not done yet).
 
I think I'm picking up what you're saying. If Apple releases an M3 Ultra Desktop at the same time as an M4 Max Laptop, customers will have equally fast computers to choose from despite one being a desktop and having the "Ultra" name. For the "Ultra" Desktop to be faster than the "Max" Laptop as the naming hierarchy implies, chips of the same generation need to be available to customers at the same time.

But as Apple's release schedule has been going, the Desktop chips are coming out so late that the Laptops have had time enough to catch up and therefore the Desktops have no advantage over the Laptops.
This... If the laptop is usually one generation ahead of the desktop, the performance of the laptop will be too darned close, especially in some years when M(n+1) is an unusually large jump over M(n), to make the desktops viable.

This year was a bit unusual, because the M3 Max added 50% more P-cores. It got a "pretty good" per-core jump, but it also has a P-core count halfway between M2 Max and M2 Ultra. That won't always be true - there will often be a generational increase in some kind of cores, but it won't always be 50% in P-cores. Some years, it may very well be 10% in GPU cores and/or an extra pair of e-cores. 50% in P-cores (or GPU cores, which we DIDN'T get this year) is HUGE!

In a "normal" year, with a modest core count jump, M(n+1) Max might fall 1/3 to 1/2 way in between M(n) Max and M(n) Ultra in performance. That's a lot more palatable than "an M3 Max is more or less an M2 Ultra on the CPU side and closer than it should be on the GPU side".

Another question will be how the RAM capacities look. If the RAM capacity for the same "size" processor continues to increase each year (and Ultra and Max continue to share memory modules - the Ultra just has twice as many), the Ultra will not generally have a compelling increase in maximum RAM. It will tend to be where it is now - 150% of the laptop, plus or minus.

If the laptops don't keep increasing in maximum RAM capacity, the Ultra may settle in at twice the laptop's capacity, which is more compelling (right now, it's reliably twice the capacity of the highest-end laptop of the same generation, BUT it's also often a generation behind, and RAM capacity has been going up with each generation).

There are two things Apple could do to increase the RAM capacity of the Ultra, either or both of which could be non-viable in a laptop. They use 8 x 16 GB RAM packages (one per channel) on the top M3 Max. 32 GB packages may well be available, but they might be too costly or power hungry for the laptop (I don't know exactly what type of RAM packaging Apple is using, but 32 GB is a pretty standard capacity for one package on a 64-bit channel). If they CAN get 32 GB packages in there, they could double the M3 Ultra's projected 256 GB capacity to 512 GB. Alternatively, many processors/chipsets can run two RAM packages per channel. There almost certainly isn't room for that in a MacBook Pro, but there could be in a desktop. Either way, they can offer 512 GB, and if they could do both at once, they might offer a terabyte of RAM to those few who needed it (I don't want to THINK about the price)...
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.