As much as I love MacPro (had a double Intel Duo Core Mac Pro, and a 5,1), I have moved on. Now an M3 Ultra Studio plus external drives perfectly satisfies my need (science calculation, image processing). The internal drive bays, memory slots, PCIe slots, and the capability to use an nVidia card in a MacPro once made lot of sense, but not any more to me. I believe many others are on the same boat.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
If they waste 5 years on a car, why not a mac pro?
- Thread starter innerproduct
- Start date
- Sort by reaction score
A halo product typically is a spiffy top-of-the line, lower unit volume product that shows off how good the product can be, with most sales being from lower end products. With iPhones the iPhone Pros are big sellers, so even though they are the top tech they are not high end niche and hence not [IMO; others may disagree] halo products.
The Mac Pro sits in the ideal position to be a classic halo product for the Macs, but it needs to exist and be really well done. ;~) The direction Apple took with its M-series silicon and Unified Memory Architecture makes building a MP in the historic modular mode challenging, but Apple's efforts with the Ultra chip configuration and 500 GB available RAM in an M3 Studio give me hope that Apple will build something as a halo Mac.
The Studio Ultra with 500 GB RAM probably would have qualified as a halo Mac product except that it was an M3 chip when M4 was already out.
Last edited:
Without expandable RAM and external GPU, the Mac pro lost a sizable chunk of its user base. It's a huge box made for a different architecture and doesn't have a very compelling reason to exist. It's kinda embarrassing
Apple Silicon is designed very differently from x86 chips. On x86 systems the CPU, GPU, RAM and other parts are separate components connected by slower links. This makes them easy to upgrade but it also adds delay, power loss, and heat. Apple Silicon instead uses a System on a Chip where the CPU, GPU, memory and media engines are all built into one package.
One big change is unified memory. Instead of separate RAM for the CPU and GPU Apple Silicon uses 1 shared pool of very fast memory. This lets data move instantly between parts of the chip without being copied back and forth. The result is better performance per watt, lower power use and much less heat. The downside is that the RAM must be chosen at purchase and cannot be upgraded later.
Another major difference is integrated graphics. On x86 systems external GPUs exist because CPUs are inefficient at graphics. Apple designed its GPUs directly into the SoC, tuned specifically for macOS, Metal and pro apps like Final Cut and Logic. For many workloads this makes external GPUs unnecessary. While this removes flexibility it also gives Apple tighter control, better reliability and fewer driver issues.
Apple Silicon also replaces many expansion cards with dedicated hardware engines built into the chip. Tasks like video encoding, decoding, AI processing and image processing are handled by fixed-function blocks that are far more efficient than general-purpose PCIe cards. This is why smaller Macs can outperform older much larger systems while using far less power.
This design explains why the Mac Pro feels awkward today. The Mac Pro was built for an older x86 model where power came from adding parts: more RAM, more GPUs, more cards. Apple Silicon works the opposite way: power comes from integration and efficiency & not expansion. That makes the Mac Studio a better fit for most professionals while the Mac Pro is left serving a much smaller group that truly needs internal PCIe cards.
It’s fair to say the Mac Pro lost flexibility that flexibility was traded for speed, efficiency, silence and reliability. Apple didn’t design Apple Silicon to replace x86 feature-for-feature. It designed it to move beyond it. That’s why the Mac Pro now feels like a product from a different era trying to live in a new one.
Yep, and I think they rolled out the current Mac Pro just to keep a promise, it really does not suit a purpose
Based on the US shipment data provided the 2023 Mac Pro M2 Ultra sells much more than the Mac Studio.
Last edited:
If the Mac Pro does stick around, I imagine that it will be pushed as a science/LLM/AI supercomputer.
As has been said many times, the Studio is now powerful enough to fit the needs of the most demanding video workflows, and to a degree even the Mac mini with a Pro chip will meet the requirements of many professionals who once needed a Mac Pro.
I think Apple will keep it because it would still provide a tidier setup than, say, 5 x Studios stacked and hooked up. That's one hell of a cable mess; plus you have the advantage of some PCIe slots for those who really want them.
As has been said many times, the Studio is now powerful enough to fit the needs of the most demanding video workflows, and to a degree even the Mac mini with a Pro chip will meet the requirements of many professionals who once needed a Mac Pro.
I think Apple will keep it because it would still provide a tidier setup than, say, 5 x Studios stacked and hooked up. That's one hell of a cable mess; plus you have the advantage of some PCIe slots for those who really want them.
That’s a fascinating idea that I’ve not seen expounded before. Makes sense, though. Would some of the data from the Apple Maps street view cars have been usable for this, too? Or the Apple Maps… walking… thingy (what do they call it?). That could have given some real-world data to assist with the “wear this thing outside and actually believe it won’t be stolen off your head” use case of giving you an AR view of the world around you, too.
That’s also a fascinating idea that I’ve not seen expounded before. Any evidence?
MacBooks account for 86% of total Mac sales; Mac desktops just 14%
MacBooks have solidified their dominance within Apple’s computer lineup, accounting for an impressive 86% of total Mac sales…
macdailynews.com
Mind you these are US shipment numbers & not worldwide.
Apple support documentation still refer to the Mac Pro as a workstation, not a desktop.
Apple still offer upgrade SSD kits for the2023 version. Even the Mac Studio Ultra versions do not have this level of support.
The rack versions have additional advantage. They offer the possibility to easily open them without having to unplug all cables.
Apple still offer upgrade SSD kits for the2023 version. Even the Mac Studio Ultra versions do not have this level of support.
The rack versions have additional advantage. They offer the possibility to easily open them without having to unplug all cables.
Mac Pro lost lot of base when they released trashcan version in 2013. iMac Pro was a band aid and 2019 Mac Pro was too late of a refresh. My last Mac Pro was 2012, and I know many folks who moved on from Mac Pro decade ago.
Yep, and when they rolled out the current model with the excessive pricing, that largely finished it off.
Total Mac revenue is a small chunk of Apple revenue. Mac Pro is a rounding number in a quarter for Apple. Mac Studio, Mini and iMac follow same Apple strategy of integrated CPU/GPU with unified RAM. Problem for Mac Pro is Apple needs to invest in R&D for expandability and they still need to figure out how to add external GPU. Too much of deviation for small stream. It also explains why Apple updates Mac Pro every few years.
Perhaps because most users don't think the present Mac Pro is a workstation, and belongs under the Arm Macs sub-forum in any case?
The problem is that they don’t have the R&D bandwidth to design both a mobile and a desktop platforms in a way that would ensure competitive pricing and software compatibility. One saving grace is that their mobile designs are faster than the mainstream desktops - so they can be sold as ultra compact workstations, but it’s not enough to outperform enthusiast x86 platform that use 3-4x more power. Frankly, I am not sure that the problem is even solvable with current technology. High-performance unified memory platforms by other key players have similar drawbacks, or are prohibitively expensive.
Some of us have hoped that Apple will have separate mobile and desktop lines, but by now it’s obvious that this is not happening. They must have determined that the ROI would be exceedingly poor. I can see folks here speculating that multi-chip-modules will change things, and I remain skeptical. Packaging is not free and I just don’t see how they can scalable build both mobile and desktop SoCs this way for cheap.
Your post is alas quite inaccurate. You are describing the traditional architecture, which is not state of the art anymore. Budget x86 platforms have been using unified memory for over a decade now (as a means to reduce costs rather than improve efficiency), and unified memory is a standard approach in contemporary supercomputing. Nvidia for example has supported unified virtual memory for many years now - a feature that Apple still lacks. And while Apple does offer unified physical memory, the bandwidth is still relatively low, and although copies can be elided under some circumstances, it does not mean that they will be always elided. Finally, we now have other vendors building AI-focused workstations with similar memory architecture but better performance and features.
All in all, the current Mac Pro has very poor value proposition, unless you are heavily invested into the ecosystem and need PCIe expandability.
Doesn’t matter if there are minor variations. Apple Mac Ro doesn’t really fit Apple silicon strategy. Apple has to solve pcie expansion and discrete GPU problem.
Let me step back and add history and context because people are mixing two very different eras of the Mac Pro.
During the Intel era the Mac Pro was built like a traditional workstation. You bought CPUs from Intel, GPUs from AMD or Nvidia, added RAM sticks, PCIe cards, RAID cards, audio cards, capture cards and sometimes external GPUs. Studios used it for film, music, science and 3D work because software and hardware evolved separately. Apple’s job then was mostly integration and industrial design.
Apple Silicon changed that model completely. Apple moved the CPU, GPU, memory and media engines onto one chip. This wasn’t about being cheap. It was about speed, power efficiency and removing bottlenecks. Unified memory here is not just “shared RAM.” It is very high-bandwidth memory tightly coupled to the chip. Designed so the CPU, GPU and AI engines see the same data instantly. That is why Apple Silicon Macs can do things Intel Macs needed much more power and cooling to do.
Because of this shift the role of the Mac Pro changed. In the Apple Silicon era most users who used to “need” a Mac Pro actually don’t anymore. A Mac Studio can now handle video, audio, coding and even 3D work that once required a tower. The Mac Pro today exists mainly for PCIe expansion: broadcast cards, specialty audio, networking and legacy workflows & not raw compute.
Apple didn’t “fail” to make the Mac Pro better. They pivoted the entire platform. The Mac Pro is no longer the center of innovation. It is the compatibility bridge for industries that still need slots.
Now about other companies that Apple used to buy parts from.
What Apple did with Apple Silicon forced the entire industry to react:
- Intel added hybrid CPUs (performance + efficiency cores).
- AMD pushed chiplets and better power efficiency.
- Nvidia leaned harder into CPU-GPU memory sharing and AI acceleration.
- ARM vendors (Qualcomm, MediaTek, others) are now building PC chips that copy Apple’s focus on performance per watt and tight integration.
Five years ago none of these was mainstream on PCs. Today it is.
Looking five years ahead I expect:
- More systems that look like “SoC-first” designs even in workstations
-Faster shared memory between CPU, GPU and AI engines
-Fewer upgradable parts but much higher baseline performance
-Expansion moving outward (network & external accelerators) instead of inside the box
No doubt that the current Mac Pro has a narrow audience and Apple knows it. That’s intentional. The future of the Mac is not modular towers but it’s highly integrated systems where most people no longer need one.
The mistake is judging the Apple Silicon Mac Pro by Intel-era rules. Apple already moved on & others are following.
And the US shipment data aligns with what you said. Laptops are fast enough for 80-86% of all users. The trans of PowerPC to Intel that started in WWDC 2005 was caused by Steve Jobs knowing that the future was mobile. 10 days from now will mark the 20th anniversary of the 1st Intel Mac, the Early 2006 Macbook Pro 15" Core Duo.
Source: https://macdailynews.com/2025/03/19/macbooks-account-for-86-of-total-mac-sales-mac-desktops-just-14/
By comparison the year before the move to Intel these were the Mac model share for 2005
- Desktop consumer Macs (iMac, eMac, mini): ~44%
- Portable Macs (iBook + PowerBook): ~43%
- Professional desktops (Power Mac/Xserve): ~13%
- Overall PC industry share for Macs was ~2–4% globally/U.S.
This reflects the approximate unit mix during a representative quarter in 2005 using Apple’s published data.
Last edited:
I feel that your analysis neglects one crucial point — while Apple Silicon raised the entry-level (and even average) capability for the end user, it also severely limited the high end. Yes, your MacBook Pro is now as fast as (and for some tasks, faster) than an entry-level desktop workstation, but if you need something more performant, you are out of luck. This is what users are concerned about when they look at the current Mac Pro and lack of high-end scalability for Apple.
RDMA does alleviate the problem to a certain extend, but only for a small class of problems and for a small group of users. It could be a path forward, but only if Apple substantially invests into interconnect technology and software layers.
Now, there is nothing wrong per se with abandoning high-end desktop. These laptops are covering a lot of use cases, and Mini and Studios are great office computers. The need for high-end local workstations is smaller than ever and most of these use cases can be competently served with a cloud solution. It's a viable business strategy for Apple, allowing it to play to their strengths. But it's gonna sting with the Mac Pro crowd. It's an emotional reaction rather than intellectual one, but it's nevertheless understandable.