Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Rickroller

macrumors regular
Original poster
May 21, 2021
114
45
Melbourne, Australia
I’m wondering if what Apple is planning for their next Big Chip is really designed to their server chip. To make money on something that’s is supposed to go into a low volume product like the Mac Pro wouldn’t make financial sense in terms of R and D, unless they already have plans to make a whole lot more of them for another purpose. The M1 Ultra is a beast of a chip already and when you look at the theoretical peak performance, it’s basically an i9 12900k + W6900X hooked up to insane memory bandwidth and speed. Would be easy enough to just add two of them for the top of the line skew and call it a day.

If anybody has any server knowledge here, how many would they need to replace whatever they have now from Intel or AMD? Considering how AMD can’t keep up with demand with their Epyc Systems, would be nice for Apple to explore this market, even if for their own internal use. They are growing their services business massively and having everything running on Apple chips could yield benefits we haven’t even thought of for them as a completely vertically integrated company.
 
Server market is more about software and Apple has very little to offer here. Maybe they have a plan, but I’m skeptical. The market is just too hard, margins too low, and Apples production volume for these chips is going to be very low as well.
 
Isn’t the low volume assumption correct if they only build them for Mac Pro? I’m not imagining that they would be doing it for anyone else’s usage. They have done it beforehand, why not now?

I reference this Seeking Alpha article from a few years back


maybe a bit too much speculation/hopium…but there are some solid points being made I think. Chips with these levels of performance and power efficiency seem tailored for the server/data center
 
All I know is it's really started to notice that many many Mac developers are NOT updating their software to at least Universal! That is beyond unacceptable today sense all but I Mac is Silicone today and your still making Intel only software! This telling me most developers are Windows Developers that found a software that will Make a Intel Mac version! That software probably doesn't write in Swift at all and don't have the skills to do it or make a universal application should have been made by now!
 
how many would they need to replace whatever they have now from Intel or AMD? Considering how AMD can’t keep up with demand with their Epyc Systems
The competition for Apple's server SOCs is Ampere Altra Max, not x64-based CPUs.
Ampere.png


Anandtech made a review of this SOC last year.
 
The competition for Apple's server SOCs is Ampere Altra Max, not x64-based CPUs.
View attachment 1972656

Anandtech made a review of this SOC last year.
Anandtech also concluded that the M series chips are no joke as it pertains to performance. I think they can compete very well against anyone at this point. Two years ago people were certain Apple was only going to be able to do low power systems maxing out at the laptop level…boy were they wrong.
 
  • Like
Reactions: Reggaenald
Isn’t the low volume assumption correct if they only build them for Mac Pro?

Apple uses cutting edge process and expensive packaging. I doubt they can reach high volume without radically changing their approach.

I’m not imagining that they would be doing it for anyone else’s usage. They have done it beforehand, why not now?

Can you elaborate what you mean by this? Like making servers for internal use only? I can see a limited utility for custom Xcode cloud servers, but that’s a significant investment for an unproven service.

The competition for Apple's server SOCs is Ampere Altra Max, not x64-based CPUs.

Uhm, why? All of these are in the same category.

Anandtech also concluded that the M series chips are no joke as it pertains to performance. I think they can compete very well against anyone at this point. Two years ago people were certain Apple was only going to be able to do low power systems maxing out at the laptop level…boy were they wrong.

Performance-wise, no argument whatsoever. And it was also clear two years ago that Apples unprecedented power efficiency can scale up to very powerful multi core clusters. But will it be economical? Apple relies expensive technologies, it is not obvious to me that the market will want to pay their fees. PC is different since Apple only covers the premium segment and their value proposition is strong there.
 
Apple uses cutting edge process and expensive packaging. I doubt they can reach high volume without radically changing their approach.



Can you elaborate what you mean by this? Like making servers for internal use only? I can see a limited utility for custom Xcode cloud servers, but that’s a significant investment for an unproven service.



Uhm, why? All of these are in the same category.



Performance-wise, no argument whatsoever. And it was also clear two years ago that Apples unprecedented power efficiency can scale up to very powerful multi core clusters. But will it be economical? Apple relies expensive technologies, it is not obvious to me that the market will want to pay their fees. PC is different since Apple only covers the premium segment and their value proposition is strong there.
I think the reason why it’s intriguing is that we are seeing what Apple has been working on in terms of silicon…but we know nothing about their software plans Which for all intents and purposes we could assume have similar trajectories.

At the moment Apple has to deal with their software and apps talking to Intel chips to provide what we now In the App Store, Apple TV+ Music and the rest. This seems suboptimal considering a billion person user base and growing. As far as volume is concerned, yes they are using cutting edge tech, we don’t know how much it really costs Apple to make their own chips. After all when you look at what they are able todo with their current hardware with regards to pricing, they are way better that what the of competition would have you believe, and I’m assuming that included a healthy profit margin per unit sold.
 
I think the reason why it’s intriguing is that we are seeing what Apple has been working on in terms of silicon…but we know nothing about their software plans Which for all intents and purposes we could assume have similar trajectories.

At the moment Apple has to deal with their software and apps talking to Intel chips to provide what we now In the App Store, Apple TV+ Music and the rest. This seems suboptimal considering a billion person user base and growing. As far as volume is concerned, yes they are using cutting edge tech, we don’t know how much it really costs Apple to make their own chips. After all when you look at what they are able todo with their current hardware with regards to pricing, they are way better that what the of competition would have you believe, and I’m assuming that included a healthy profit margin per unit sold.

The expensive part is probably not the chips themselves but the packaging: RAM, wide interfaces, all that stuff. Those RAM modules are crazy expensive in their own, and Apple uses some sort of custom to order stuff.

On the software side there is no indication that they plan anything particular. Servers run on Linux after all and nothing is going to change in this regard. Apple could improve virtualization, which would make running Linux VMs as close to iron as possible viable.
 
What differentiates Apple’s MacOS and iOS software is ease of use for the casual user and creative professional. Servers tend to be used by highly trained sysadmins who look to use a complex stack of software, and this is an area where Linux holds the advantage.

So a lot of Apple’s advantage of vertically integrated hardware and software doesn’t really apply. Thats why I would be surprised to see Apple making another push into server hardware, where the real money is in large-scale operations such as cloud providers. It just seems to take focus away from Apple’s to,date very successful business model.

Early indications are that Mac Studio and the M1 Ultra are a strong value proposition in the workstation market, competing with setups that sell for double the cost in the PC market.
 
I guess we will have to wait and see. Still to be in this era of Apple silicon advancement is a privilege. Also I would think the big chip will be introduced before the M2 cycle, which would give buyers a nice signal in regards to product cadence.
 
  • Like
Reactions: saintmac
Apple's SoC architecture is not a good match for cloud servers. Apple makes balanced systems, while servers are often customized for a specific purpose. For example:

RAM: Cloud servers have 256 GB to 24 TB memory these days. A system with 2x M1 Ultra would only be suitable for tasks where the user does not care about memory.

SSD: Systems with 8 internal SSDs are common in applications where the user needs fast local storage. Each of them is comparable to the SSDs used in the Mac Studio.

GPU: Either the GPU is a completely unnecessary waste of money, or the user is going to need a lot of GPU power. 8x Nvidia A100, for example.
 
It would make sense for Apple to release servers for Small business/companies.

A Server that preserve your company's data & privacy locally instead of using data harvesting cloud hosting.

And the software support is right there, as MacOs can run almost anything made for Linux with little tweaking.
 
If you want to run something as a local server, Apple has you covered. I wouldn't be surprised if they were developing hardware to run their own services, but Apple isn't really set up to provide the type and level of support required for servers. See also: Xserve.
 
It would make sense for Apple to release servers for Small business/companies.

A Server that preserve your company's data & privacy locally instead of using data harvesting cloud hosting.

And the software support is right there, as MacOs can run almost anything made for Linux with little tweaking.

I would instantly buy an Apple NAS.
 
It would make sense for Apple to release servers for Small business/companies.
A home/small business server could make sense. The only problem is that Apple is treating internal storage like a luxury rather than a basic feature. Release a Mac Mini with two NVMe slots for additional storage, and I would replace my QNAP NAS with it. At least if I don't have to pay extra for something ridiculous like the M1 Pro, because the M1 is already too powerful for such servers.
 
If you want to run something as a local server, Apple has you covered. I wouldn't be surprised if they were developing hardware to run their own services, but Apple isn't really set up to provide the type and level of support required for servers. See also: Xserve.

Mac Server app used to offer full-fledged server functionality but in the recent years it has been reduced to profile management only. Some components like file/backup server has been integrated into the base OS, some, like web server/wiki server/LDAP have been dropped entirely as far as I know
 
A home/small business server could make sense. The only problem is that Apple is treating internal storage like a luxury rather than a basic feature. Release a Mac Mini with two NVMe slots for additional storage, and I would replace my QNAP NAS with it. At least if I don't have to pay extra for something ridiculous like the M1 Pro, because the M1 is already too powerful for such servers.

It would need to have hot-swappable SATA and NvME slots with solid RAID support.
 
data harvesting cloud hosting
What are talking about? Any example?

Uhm, why? All of these are in the same category.
I don't think they compete with each other. People choose x64-based virtual machines in the cloud because they can't choose ARM-based virtual machines. ARM-based instances are much cheaper than x64-based instances on AWS.

home/small business server could make sense.
Xserve predates AWS, so it made more sense then. What company today buys a server instead of using a cloud service provider?
 
I don't think they compete with each other. People choose x64-based virtual machines in the cloud because they can't choose ARM-based virtual machines. ARM-based instances are much cheaper than x64-based instances on AWS.

It depends. For many applications it really doesn’t matter as software can compile to either target. If I were looking for a cloud host for my custom backend software I would choose ARM for price. If you are running a prepackaged x86 docker image, sure, you need an x86 host.
 
Xserve predates AWS, so it made more sense then. What company today buys a server instead of using a cloud service provider?
Local file servers are common, because they are faster and cheaper than the cloud. Storing the data in the cloud is cheap, but downloading it typically costs something like $50 to $100 per terabyte.
 
  • Like
Reactions: Nermal
Xserve predates AWS, so it made more sense then. What company today buys a server instead of using a cloud service provider?

For file storage? Who is using cloud storage for files? It’s expensive and very slow. Cloud storage in this context is usually delegated to backups.
 
A revived Apple server sounds interesting. There's no question the hardware chops could be there. But Apple would need to bone up its software in an astronomical way to even put a slight ding in the server universe. macOS Server may still be a thing, but it straddles a line between being deprecated & proof of concept.

A glaring example is macOS Server's "Profile Manager" service. It's not robust; it will flail after about 50 users, the database will get corrupted, and you're SOL. You should probably have invested in JAMF or any number of the MDM providers out there instead of mucking around with macOS Server's Profile Manager.

And let's be real: it's not going to be hosting webpages. Which gets back to the point. Where would this fit in, in today's environment?

Loads of companies & organizations moved to remote during the pandemic. Many are still operating this way & they're not going back to full-in-office. There may be the nostalgia for the Mac server but as it stands now it would have to have to bridge to Apple's cloud services to be remotely useful; kind of along the lines of the caching server concept.

For example, let's say Apple built a M1 Ultra-based local server that could be managed through iCloud. It does its identity & cloud storage, etc. all on Apple's cloud servers. But has a 2-way trust with your M1 Ultra on-prem. It mirrors all of that cloud content back & forth to your M1 Ultra server on-prem, so everything's super fast on your network when you're there. Those files get mirrored to the cloud, if desired, so accessing these files & services are seamless for the user without having to muck around with VPNs.

Thoughts?

I just don't think Apple has a desire to get back into selling Mac servers again.
 
What would be the specifications of a modern Xserve assuming a similar price to the old Xserve (US$3000 - US$4000)?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.