MP 7,1 7,1 Bros: We're headed into 2025. How are we feeling?

[Norbert Mikołajczyk]

I've been a Mac Pro user since 2010, and while I love my 2019 model, I can't help but feel abandoned by Apple's poor decisions—especially for a so-called professional workstation.

What really pushed me over the edge was the lack of GPU support for the recent and current Mac Pro lineup. Not only is that a stupid decision, but it's outright insulting. And let's not even get started on Sequoia and the lackluster support for a $5,000 W6800X Duo MPX.

Once I'm done with my 2019 Mac Pro, I'm leaving this platform for good. It's no longer worth the money, effort, or time. Apple had absolutely nothing to lose by supporting at least the AMD 7000 series, but they chose not to—so I’m choosing not to give them my money anymore. **** Them.

I hear you! It’s more than this for me. Every Apple desktop product is fully soldered, no way to upgrade even ram, ssd. And those horrendous prices!
I sold my macintel macbook pro recently, now i'm selling MP.

The warning shot for me was recent message from blender community that they are stopping future updates for macintel because of old and buggy macOS drivers for amd.

I had opportunity on the rtx 5090 launch and bought a small, custom watercooled pc with ryzen 7950x3d and rtx 4090 in very small formd t1 case with the bag.
I'm very pleased with it, it's not as silent or power efficient but who cares. It is still great, have the same/similar single core cpu performance and better multicore than apple silicon and this gpu is so powerful!
I loaded some small apps for win10 to have the functions I liked on Macos, for example drivers and software for the Apple touchpad for mac gestures like 3 finger drag.
And all that for 1/3 of a macstudio price.
Oh did I mention I can upgrade all the parts?

 

[deconstruct60]

Apple had absolutely nothing to lose by supporting at least the AMD 7000 series, but they chose not to—so I’m choosing not to give them my money anymore. **** Them.

Nothing to loose but money.

No new Intel systems means that the number of target MP 2019 they could place into is static (and shrinking over time).

Unless the drivers are extremely specific the AMD GPU packages used by Apple, not long after the Apple cards came out the drivers would be exploited to be used with Off the shelf Windows cards. Again shrinking market of the target cards. Each Windows card sold goes toward paying Windows/Linux drivers. Apple is extremely likely the only one collecting money to pay for the drivers. For example, 50-60% folks only paying the Windows fee would leave a substantively large shortfall.

If Apple was selling next gen MP with W7000 they could skip the fees all of them, but likely were not planning to.

So where is the large user base paying for the work to be done?

Nevermind that a large fraction of the money to pay for next gen AMD GPUs come out of non-MacPro sales ( laptops with dGPUs sold in substantially higher numbers, large scrren iMacs also. ). The 'danger' is when the whole rest of the Mac line up leaves the dGPU space. The MP by itself is not volume units. Shrink that to already sold MPs and that is even smaller.

When all those other folks who were helping to pay leave; it is a problem.

 

[avro707]

Based on the above, people are more likely to not spend five figure price tag on a locked down silicon Mac Pro and more likely to to spend big $$$ on a windows workstation which can be readily upgraded as the need occurs.

We can see that already with all the folks who have moved over already. If it keeps going then the Mac Pro no longer needs to be offered. And maybe the Mac Studio also needs to be dropped.

 

[deconstruct60]

We can see that already with all the folks who have moved over already. If it keeps going then the Mac Pro no longer needs to be offered. And maybe the Mac Studio also needs to be dropped.

I highly doubt Apple thought they were going to maintain the same number of buyers for the MP 2019 when they increased the entry price by 100%. The Nvidia super fans ... again doubtful that was a 'surprise' and made before the move to Apple Silicon. The product already has a "low volume" tax on it.

Some folks are going to leave , but Apple doesn't have to sell to the exact same set of customers who bought previous generations of Mac Pro's. If get new customers to replace old ones , then the user base shrink won't be as bad. That is more on software than hardware. The incremental speed increases from M1 , M2 , M3 , M4 has been decent. The Mac Pro isn't going to get yearly so if jump 2-3 gens at a time the bumps are even more decent.
What is more so missing so far is substantial software to fill more Mac Pro seats with new users. Some at the OS level ( infrastructure quirky problems ) and some at application level ( better leverage of unified memory across compute units).


The Studio is largely based around the Max and Ultra SoCs. The Max SoC is going no where. The MBP 14/16" will keep that healthy. So if Apple already has it why wouldn't they use it in a Studio? The Studio is a replacement for the large screen iMac. That is largely independent of the Mac Pro in the line up.

Throw on top that Apple recently made the Mini Pro smaller ... it is boxed in thermally. The Studio with better thermals isn't likely going anywhere. ( The Mini enclosure even more so excludes the Max SoC. )


As for the Ultra chip. UltraFusion works. If just two Max chips ... the Max is just going to be around. Most likely Apple will be doing more with multiple chip packaging over the long term than cancelling it altogether. ( Just different trend lines for I/O and logic on density increases and wafer costs steadily increasing. And advanced packaging costs likely coming down a bit with scale and tech refinements. ). Building two die Ultras should get easier over the next 3-4 years. If squeeze out some of the Ultra cost that will leave the Studio in all the more healthier position.

It would likely help the Mac Pro if the Ultra was build from something other than a very chunky chiplet. If that becomes a better partitioned chiplet components in the future the Ultra would be on safer ground. ( And probably also the Mac Pro. Something incrementally (not necessarily doubling) bigger than a Ultra would help the Mac Pro (the quad thing doesn't make economic sense or technical in several ways).

Is the Mac Pro going to get back to era of yearly upgrades ? Highly unlikely. Two years ...maybe not. But 3 year cycles ... would keep folks in the dark for a long time before the schedule becomes apparent.

 

[bcomer]

I picked up a basically new M2 Ultra (128GB/4TB) for a very fair price.

The PCI card mounted SSDs have been moved to OWC TB4 Express 1M2 housings.

This is so much faster than my 2019 MP (12C/1TB/192GB).

 

[Polymorphic]

I picked up a basically new M2 Ultra (128GB/4TB) for a very fair price.

The PCI card mounted SSDs have been moved to OWC TB4 Express 1M2 housings.

This is so much faster than my 2019 MP (12C/1TB/192GB).

I’m leaning that way. Already picked up the Express 1M2 housings, so now I’m just trying to decide whether to wait for the M4 Studio.

 

[bcomer]

I’m leaning that way. Already picked up the Express 1M2 housings, so now I’m just trying to decide whether to wait for the M4 Studio.

For the money involved I needed to choose between the M2U with 128GB/4TB or a new M4M with 64GB/1 or 2TB. I went with the M2U as stated. The M4 Studio would have been more money.

Buying the Studio used but basically new condition just made better sense. Added bonus was no sales tax.

 

[SDAVE]

Hear hear!

I don’t trust what Apple might do. So I would wait and see what they actually do and how that goes.

Apple could have quite easily worked with AMD to do drivers for the 7000 series but chose not to. AMD reps pointed the finger at Apple (I have their email).

Looking at Nvidia 5090 now it is released that’s not an option either with exorbitant prices and scalpers trying to profit. Not to mention the power cable issue that some folk have raised.

I should have taken out a loan and purchased 10 of them and sold them at double price.

The AMD 9000 series was just announced

Would be amazing if apple adds support but i doubt it haah

 

[Regulus67]

The AMD 9000 series was just announced

Would be amazing if apple adds support but i doubt it haah

Can't let an old machine compete slightly with the shiny brand new macs. Apple ™
Though it looks very good

 

[avro707]

The AMD 9000 series was just announced

Would be amazing if apple adds support but i doubt it haah

Interesting prices:

AMD Unveils Next-Generation AMD RDNA™ 4 Architecture with the Launch of AMD Radeon™ RX 9000 Series Graphics Cards

– The new AMD Radeon™ RX 9000 Series graphics cards deliver enthusiast-level gaming experiences supercharged by AI –

www.amd.com

Not sure about the 16GB memory.

I'm sure when they actually become available people will be selling them at USD$10,000 each.

 

[Norbert Mikołajczyk]

Interesting prices:

AMD Unveils Next-Generation AMD RDNA™ 4 Architecture with the Launch of AMD Radeon™ RX 9000 Series Graphics Cards

– The new AMD Radeon™ RX 9000 Series graphics cards deliver enthusiast-level gaming experiences supercharged by AI –

www.amd.com

I'm sure when they actually become available people will be selling them at USD$10,000 each.

Yeah, the stock will be limited, scalpers and trumps tariffs will do the rest

 

[deconstruct60]

Yeah, the stock will be limited, scalpers and trumps tariffs will do the rest

AMD supply likely isn't going to be anything like the Nvidia availability for their 5080/5090. AMD doesn't really have reference cards so the prices for most cards is likely going to be higher. That will slow some of buying. Super hardcore Nvidia fanboys won't buy anything else.

If everyone desperate for a card charges into the 9070XT then it will go fast. But that would be loads of 'defectors' driving that. The 'plain' 9070 is the same base chip and VRAM so there is almost no incentive to provision lots of those early on. Even more so when Nvidia has even less supply.

Even in best conditions there likely isn't enough initial supply to stop scalpers to wiping initial cards off the table if the hype train is high enough for a possible temporary shortage. Apple comes up short on iPhones for some launches.

Tariffs at 10% would put those at about $599 and $649 instead of what AMD has in the initial press release.

Intel has some 'surprisingly' not so competitive CPUs so AMD is assigning more wafers to exploit that. Nvidia has a 'print money' machine on data center GPUs and have assigned more wafers to that. Similarily for Apple the rest of the Mac line up has adjusted very well to the M-series transition. That is where the investments are skewed too. (not hardware that was initially sold 4-5 years ago. )

 

[Regulus67]

The AMD cards do look pretty good. And there will most likely not be any issue with availability, it seems.

Too bad we can't use them in macOS

 

[impulse462]

Now that the M3 Ultra isn't coming to the mac pro, its probably the end. With the 7,1 i guess it was just a stopgap between an intel mac workstation and the take over of apple silicon mac studio. mac studio is a fine product but it is not a classic workstation where you can mix and match internal components. its clearly obvious this is no longer a market for apple.

as for my 7,1, I can't say i haven't gotten good use out of it. I basically did almost all my experiments for my entire phd thesis on it since it is running linux w/ 2 nvidia gpus. i think that was worth the money even though i clearly could have built a much cheaper and more powerful workstation. ive always wanted a mac pro since my parents didnt get me one when i was younger. its still running rock solid so I have no complaints.

as for next hardware though, i most likely will not be purchasing a mac desktop in the future. for me not being able to swap internal components means the computer isn't a desktop. this might be an old way of thinking in 2025, but i'm standing by it. i look forward to continuing to use my 7,1 as a linux box and building a more powerful workstation for more research in the future.

 

[Norbert Mikołajczyk]

Yeah, Apple Silicon Mac Pro looks dead too. I sold my last Intel Mac this week because I do not need such a powerful expandable Mac anymore as I changed my job last year.
So yeah if you want and need an expandable and upgradable, the pc is the only way.

For me the Mac Studio is too big, too pricey and not upgradable but the shared 512gb memory is amazing for ai, so it will be awesome machine for some users.

Ps. I’ve just bought Mac Mini M4 Pro on a sale to accompany my pc. The small form factor is amazing!

 

[Regulus67]

I prefer macOS.
Even if I also use win10 for gaming. Perhaps I'll use SteamOS instead, if the rumours are correct.
So I am still very happy with my 7,1s.

I am also considering upgrading to 768GB RAM on one machine. Because I have a large text archive, with over 14 000 pages. If it is possible to train the DeepSeek to help me find texts in the archive, and give answers based on that.
It is a long and tedious process to search it manually. Which I have tried several times during the past 20 years 😆

 

[deconstruct60]

Now that the M3 Ultra isn't coming to the mac pro, its probably the end.

The M4 Max in the MBP didn't release at the same time the M4 Max in the Mac Studio did. ( In fact for all of thee MBP 14/16" intros on M-series, the Studio hasn't lined up with any of them. ) The Mini didn't even get a M1 Pro.

There is basically zero evidence that the Mac Pro won't get a M3 Ultra later. There is zero technical requirement that the Mac Pro update at the exact same time as the Mac Studio. Two different products in two different containers aimed at two substantively different markets.

If Apple is sucking a fair number of M3 Ultras into their data centers, there could be not enough to go around. A major factor is what the M3 Ultra production line is capable of.

Making the Mac Pro look the most 'sexy' offering all of the time is not the most primary factor at play . The mac Pro update is more likely not going to be 'timely' rather than not done at all. Two year update schedule would be 'aggressive' given Apple's past decade of Mac Pro updates. It was delusional to think that moving to Apple Silicon was going to get magical , yearly updates for the Mac Pro. It wasn't components slowing the update cycles down ... it was Apple ( and buying cycle frequency rates for this submarket) . That doesn't have diddly squat to do with who made/makes the CPU package.

When people generally buy at slower pace the updates come at slower pace. That's it.

 

[Boil]

The M4 Max in the MBP didn't release at the same time the M4 Max in the Mac Studio did. ( In fact for all of thee MBP 14/16" intros on M-series, the Studio hasn't lined up with any of them. ) The Mini didn't even get a M1 Pro.

There is basically zero evidence that the Mac Pro won't get a M3 Ultra later. There is zero technical requirement that the Mac Pro update at the exact same time as the Mac Studio. Two different products in two different containers aimed at two substantively different markets.

If Apple is sucking a fair number of M3 Ultras into their data centers, there could be not enough to go around. A major factor is what the M3 Ultra production line is capable of.

Making the Mac Pro look the most 'sexy' offering all of the time is not the most primary factor at play . The mac Pro update is more likely not going to be 'timely' rather than not done at all. Two year update schedule would be 'aggressive' given Apple's past decade of Mac Pro updates. It was delusional to think that moving to Apple Silicon was going to get magical , yearly updates for the Mac Pro. It wasn't components slowing the update cycles down ... it was Apple ( and buying cycle frequency rates for this submarket) . That doesn't have diddly squat to do with who made/makes the CPU package.

When people generally buy at slower pace the updates come at slower pace. That's it.

WWDC 2025, Apple releases a redesigned Mac Pro with M5 Ultra/Extreme chips in the smaller chassis Gurman has talked about in the past...?

Alongside the all-new M5 Extreme Mac Pro Cube, of course...! ;^p

 

[H. Flower]

I agree its not entirely clear if the Mac Pro is cancelled. One theory is they are giving it M5 extreme later this year to differentiate it from the Studio.

Or .... maybe it is discontinued.

Once again, Apple has left me in limbo. It'd be great if they could share their roadmap. Lol !

 

[reeltime100]

I agree its not entirely clear if the Mac Pro is cancelled. One theory is they are giving it M5 extreme later this year to differentiate it from the Studio.

Or .... maybe it is discontinued.

Once again, Apple has left me in limbo. It'd be great if they could share their roadmap. Lol !

Seems to me the logical conclusion is there's a M4 Ultra that will be Mac Pro only. They left that room in the lineup, and they were killed online for not giving the M2 Mac Pro a reason to exist. Giving it their top-dog processor would make the most sense.

 

[deconstruct60]

WWDC 2025, Apple releases a redesigned Mac Pro with M5 Ultra/Extreme chips in the smaller chassis Gurman has talked about in the past...?

Way cheaper for Apple to release M3 Ultra in exact same container that they have essentially already paid for. (and a SoC that they need volume in a revenue generating product to help pay for ).

If the M5 Ultra is using a modified die of the laptop M5 Max then the likelihood of that one shipping before the laptop version is very low. The laptop version can practically pay for itself. The Ultra subcomponent is likely on vastly thinner ice. Pretty good chance Apple is going to ship what makes money first.

One of the primary problems Apple has with the Mac Pro is consistent updates. Not the most 'sexy' update at random times , but consistent delivery. A Mac Pro on a regular 2 year update sequence would help far, far , far more than some kind of ego pissing match product that has to have some 'dominating' version number slapped on it that is only fleeting anyway. ( a M6 would likely ship before any M5 Ultra comes along. The single thread crown would go to that before the Mac Pro updated anyway. )

Apple giving forward guidance that there is likely no "ultra 4" is more likely to ground expectations that M3 Ultra is what you'll get in the immediate future. Not some hyper acceleration to M5 .


The smaller chassis doesn't make much sense. The Mac Pro is sold both tower and rack. Racks aren't getting narrower so making the tower shorter buys what?

Fewer slots, but still no Nvidia/AMD GPUs ... that isn't going to buy those fanboys back at all. Meanwhile squeezing out the folks that have 4-5 older PCI-e (and PCI tech) cards that scale out on slots. Those folks would leave also.

Chopping away slots because Apple isn't putting substantive 'value add' for the slots is completely missing the root cause of the problem.

Alongside the all-new M5 Extreme Mac Pro Cube, of course...! ;^p

Mac Studio M4 Max is a Cube killer. . Studio M5 Max will be an even bigger Cube killer.

 

[deconstruct60]

Seems to me the logical conclusion is there's a M4 Ultra that will be Mac Pro only.

Hardly a logical conclusion when Apple is , unusually, overtly hinting that there is no M4 Ultra.

"... A spokesperson for Apple has told French technology website Numerama that its M4 Max chip lacks an UltraFusion connector, ..."
https://www.macrumors.com/2025/03/05/apple-confirms-m4-max-lacks-ultrafusion/" ...

Apple typically doesn't talk about future products or non products. More than decent chance Apple is trying to nip 'M4 Ultra just around the corner' speculation in the bud with this reveal.

What Mac Pro fans want to hear is that Apple is going to kneecap the Mac Studio so the Mac Pro gains more traction. That is likely mostly just 'wishful thinking' ; not logic.

They left that room in the lineup,

It isn't leaving room in the line up if there is no M4 Ultra. That option isn't in the line up.

and they were killed online for not giving the M2 Mac Pro a reason to exist. Giving it their top-dog processor would make the most sense.

Giving the Mac Pro a processor that can be paid for would make the most sense. Assigning a narrow fringe processor that the Mac Pro revenues probably can't pay for doesn't really make the most sense.

Pretty good chance that Apple is going to target relatively narrow niches with the ultra studio and ultra Mac Pro. The primary, "general workstation" target is the M4 Max Studio. That SoC is far easier to make and the system is far more affordable to a wider set of users. If Apple sells substantively more of those, then that will be fine with them.

At some point Apple will trot out a M5 Ultra to this smaller niche. Perhaps when primary Studio could go to M6 Max and the M3 Ultra just got too deep on the wrong side of the mismatch. It would help if it was a bit sooner than that though.

Unlikely, that Apple is going to go to some yearly Ultra cadence. Apple has set the expectations that the Mac Pro isn't a yearly update product.


The two major problems the Mac Pro has at the moment are

1. inconsistent delivery. Pick a 2 (or worse 3 ) year schedule and stick to it. The Mac Pro is chronoically a Rip van Winkle product that is comatose for long periods of time. If what they have in 2025 is a M3 Ultra , then it would make the most long term sense for Apple to get a MP 2025 out the door . Setting the pace is very important; not bragging rights for a few.


2. software kneecapped more than hardware.
The scope of the driver migration to modern foundation isn't extremely healthy. ( Flakey storage cards are a show stopper issue. Apple's attitude that only their SSD matters is a highly limiting factor. ).

It won't matter if have a Mn+1 SoC in there and the drivers for the cards are flakey. (e.g., Intel GPU drivers in year one. )
Half baked Apple Intelligence is mainly a software issue; not a hardware one.

( very similar issues to the iPad where the SoC isn't really limiting anything substantive. The OS is lagging behind.)





P.S. There is a chance Apple could so a M4 "bigger than a Max" that wasn't a GPU core doubling. It would be somewhat confusing to label that Ultra ( if intended to use Ultra to cover muliple chip packages in future). It would give them something 'bigger than a Max' to offer that just used a single die.

 

[Regulus67]

Hardly a logical conclusion when Apple is , unusually, overtly hinting that there is no M4 Ultra.
...
Apple typically doesn't talk about future products or non products. More than decent chance Apple is trying to nip 'M4 Ultra just around the corner' speculation in the bud with this reveal.

I agree. And the M3 Ultra has the new Thunderbolt 5, so an updated chip.
Some of us believed Apple would present a new Mac Pro with Thunderbolt 5, along with an updated Pro XDR display. Because of the higher bandwidth it offers.

Looking back to an unusual comment made by Apple in WWDC 2020. Where "one more Intel mac is in the pipeline".
Only very recently did I understand that it was the iMac 27" 2020, which was indicated. As I missed that it also got the T2 chip and Apple's soldered SSD.

 

[impulse462]

The M4 Max in the MBP didn't release at the same time the M4 Max in the Mac Studio did. ( In fact for all of thee MBP 14/16" intros on M-series, the Studio hasn't lined up with any of them. ) The Mini didn't even get a M1 Pro.

There is basically zero evidence that the Mac Pro won't get a M3 Ultra later. There is zero technical requirement that the Mac Pro update at the exact same time as the Mac Studio. Two different products in two different containers aimed at two substantively different markets.

If Apple is sucking a fair number of M3 Ultras into their data centers, there could be not enough to go around. A major factor is what the M3 Ultra production line is capable of.

Making the Mac Pro look the most 'sexy' offering all of the time is not the most primary factor at play . The mac Pro update is more likely not going to be 'timely' rather than not done at all. Two year update schedule would be 'aggressive' given Apple's past decade of Mac Pro updates. It was delusional to think that moving to Apple Silicon was going to get magical , yearly updates for the Mac Pro. It wasn't components slowing the update cycles down ... it was Apple ( and buying cycle frequency rates for this submarket) . That doesn't have diddly squat to do with who made/makes the CPU package.

When people generally buy at slower pace the updates come at slower pace. That's it.

On the surface I get your point, but I still don't really buy it. Is economies of scale really an issue at this point? Their supply chain has been /mastered/ at this point, i don't see any reason to not update it in the mac pro unless they arent planning to or its getting a newer variant.

 

[Mac3Duser]

The main problem is what is a mac pro for and how many years does this kind of workstation last.
The workstation is the pcie slots. Whether the Xeon is fast or not is not the problem, whether it is overtaken by a modest latest generation i5 in benches is not the problem, the problem with the workstation is the number of pcie slots.
but not only for storage.
For graphics cards that can be changed as needed, for wifi, raid, ethernet, sound cards, and everything else you want with the right drivers, depending on the projects.
the workstation must therefore be very open to adapt.
how many years is a workstation valid? Five years for a use that requires a lot of power, but almost ten years in fact for a more open use, since hardware upgrades are possible, and each year costs less (more ram, more powerful processor, storage that you change, gpu that you change etc...) as long as the pcie slots work.
Apple's actual Mac Pro doesn't fit into this logic at all:
- no RAM upgrade possible
- no GPU additions
- only storage, ok
so it's a very expensive computer for a few years of use.
If Apple wants to exist in the workstation market, it needs to offer card addition possibilities, all cards, and without constraints, or create a modular workstation where you can add as many processor cards as you want: up to 8 M4 cards, or 8 M4 pro cards or 8 M4 cards max, with a power supply for such configurations, and PCIe lines for storage, wifi, ethernet, sound...
whether in rack or towers, such a modular chassis would be interesting for the workstation, server and AI computing market.