Or it could be that Apple, will have 12 in hardware, but only 6GB available unless you pay the Apple tax.
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.
Apple M5 Chip's Dual-Use Design Will Power Future Macs and AI Servers
- Thread starter MacRumors
- Start date
- Sort by reaction score
yknow, a true core memory of mine on here is one of the top comments in the original iPod announcement thread being about the need for Apple to focus on making its Xserve line better. wild to me that so many people don't grasp how Mac mini has become their de-facto server product at a fraction of the cost.
incredibly exciting stuff out of both Apple and TSMC. this is the kind of rapid growth I was looking forward to greatly when Apple Silicon was first announced; Apple free from the whims of Intel is a whole other beast.
on the flipside: hey Tim, since yall fumbled the bag *so* hard on this that only the latest, most maxed-out iPhone can run Apple Intelligence—can we get 16GB RAM default on all Macs now?
on the flipside: hey Tim, since yall fumbled the bag *so* hard on this that only the latest, most maxed-out iPhone can run Apple Intelligence—can we get 16GB RAM default on all Macs now?
M5? I'm still using M1 Mac Mini while holding out for M4 Studio. I'll never catch up. 🙁
No, this is so Apple won't have to rely on Nvidia for the chips. Apple can optimise the chips for their needs and probably save some money and time.
The Nvidia chips are very expensive and getting them may take a long time unless you pay extra. Also the relationship between Apple and Nvidia hasn't been good for the last 10 years.
in 2007 Apple Computer, Inc became Apple, Inc.
Yet today, Apple is more of a “computer” company than they ever were before the name change. Today they design their own CPUs/GPUs, OS, programming languages, developer tools, and now cloud AI servers.
Two big things they do not do themselves are R&D for new chip fabrication technology and manufacturing/assembly. They seem pretty steadfast in their avoidance of manufacturing, but I wonder if they might get into fab process R&D. It might be appealing to them to own some special fab sauce that nobody else has.
Yet today, Apple is more of a “computer” company than they ever were before the name change. Today they design their own CPUs/GPUs, OS, programming languages, developer tools, and now cloud AI servers.
Two big things they do not do themselves are R&D for new chip fabrication technology and manufacturing/assembly. They seem pretty steadfast in their avoidance of manufacturing, but I wonder if they might get into fab process R&D. It might be appealing to them to own some special fab sauce that nobody else has.
Give it another 5 years. Tim Apple might retire by then.
My Mac mini M2 Pro with 32GB ram, 2TB SSD, 10GB ethernet, 12 core CPU, 19 core GPU is a very competent file server with attached RAID drives. I do have a 10GB ethernet network so transfers of data are at the fastest rate currently available to the public. That is one heck of server for $2,429 (military discount) plus sales tax and Apple Care.
Siri, summarize this article:
“I’m going to be be be be be be be be smarter in a few years “
“I’m going to be be be be be be be be smarter in a few years “
What does it mean for users, in terms of gap between m1-m3 and this new m5? Will there be a huge jump ?
I'm still just blown away they are charging one THOUSAND dollars just for your RAM & SSD upgrades there
(At retail prices, that spec goes from $1699 to $2699, just for the RAM/SSD upgrades)
So will base-model M5 Macs finally come with adequate amounts of RAM, or will the $1299 M5 iMac still have less memory than a $799 Android phone? 
Or they plan on using the chips themselves.
The mac desktop lineup is updated as (less) frequently as with intel, so…
For that price one might as well buy the studio. Which, of course, is the whole point of apple’s memory and ssd upgrading price scheme.
And it’s smart, NVIDIA and the rest of the AI bubble will have popped by then. That will leave only *profitable* AI businesses intact.
Apple is wholly immune because they weren’t dumb enough to jump on the NVIDIA hype train for hardware. They also are approaching AI in a completely different context (read: useful to *regular people* in their daily lives) than the entire industry.
I suspect we will get N3P/X for M5 as the cost for N2 will probably again increase significantly and TSMC are only ramping the process up in 2025 (last thing I read was high-volume production in H2 2025 - Anandtech).
16GB RAM as base configuration in the Macs would certainly be appreciated, though (to me it feels like this is overdue).
Apple is a direct development partner to TSMC. They literally are the only reason N3 (by way of N3B) was feasible to even produce after everyone else pulled out. Apple’s capital investment in TSMC’s lines make it almost a certainty that they’ll be first in line for each new node because the nodes themselves aren’t economically viable without massive upfront capital investment.
Dual use design means that the components Apple designs can be used for multiple applications. It’s really not specific to the Ultra series.
The bigger point of what Apple might start doing with the M5 series is the use of SoIC packaging technology. Currently Apple uses monolithic SoC dies, meaning a variety of system components are crammed into a single square of silicon (CPU, GPU, NPU, memory controller, Secure Enclave, display controllers, USB/Thunderbolt controllers). For smaller chips like the base M this works fine, but as you start to scale it up the SoC dies get to be quite large. The Max is about as large as is practical to make a single die, hence why the Ultra is basically two Max chips stitched together. But this isn’t ideal because you’re wasting silicon space on those extra controllers you might not need (you don’t need two Secure Enclaves for example).
SoIC means Apple can fabricate all of those SoC components separately and merge them together into a smaller and more efficient package (that performs better than a monolithic die). It also means Apple can more easily customize its chips rather than having to tape out an entirely new die. This is where the “dual use design” comes in. Apple can take CPU, GPU, and NPU tile designs from its consumer chips and create custom packages that are ideal for AI servers (for example lots of CPU and NPU cores but skimping on GPU). Because everything is modular there’s more flexibility when it comes to custom use cases.
Also in 2007, Apple Computer Inc. settled their long-running trademark dispute with Apple Corp. (the Beatles' record label) which freed them up to simplify the name of the company.
No wild theories about Apple moving away from their computing roots are required.