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

MacRumors

macrumors bot
Original poster
Apr 12, 2001
70,131
41,688


Apple doesn't release an "Ultra" variant for every Apple silicon chip, but the company is planning to debut an M5 Ultra chip in 2026, reports Bloomberg.

mac-studio-green.jpg

The M5 Ultra is slated for the Mac Studio, and it's also likely that Apple will use it in a Mac Pro update. There's no word on when the M5 Ultra might come out, but historically, Ultra chips have followed the Pro and Max chip variants.

Apple already debuted the M5 chip in the MacBook Pro, but we'll be waiting until early 2026 for the M5 Pro and the M5 Max. The M5 Ultra will likely come after that, perhaps in June or even September 2026.

Earlier this year, Apple decided to launch an updated Mac Studio with M4 Max and M3 Ultra chip options, making it clear there was no M4 Ultra. Ultra chips are two Max chips fused together with "UltraFusion" technology, but the M4 Max chip does not have the UltraFusion connector needed for an M4 Ultra chip.

Apple also made it clear that it did not plan to create a high-end Ultra chip for every M-series generation.

There are no other rumors of new features for the Mac Studio, so the M5 Ultra update might be focused on internal improvements. Apple is rumored to be developing a pair of new displays, and at least one of those could launch alongside the Mac Studio in 2026.

Article Link: M5 Ultra Chip Coming to Mac Studio in 2026
 
New Apple displays too then in 2026?


M5 Max has not yet been released.
Ah, so it’s the Max variant that has the connector, hypothetically. The M4 Max didn’t have the connector. I thought the connector was part of the base processor architecture.
 
Ah, so it’s the Max variant that has the connector, hypothetically. The M4 Max didn’t have the connector. I thought the connector was part of the base processor architecture.
Only the Max variants have had the connector if I remember correctly. The Base, Pro, and Max are all different silicon, and it's just the Max variant that gets doubled up with the connector to double everything.
 
Waiting for M6 2nm will be good new packaging new architecture this current architecture already same from A16
 
This will be the most powerful AI machine…

…that doesn’t run AI.
lol what

pytorch runs fine

Local models work fairly well despite the drawbacks of M4 (low inference speed mostly).

M5 will hugely improve this: per GPU core neural acceleration, matrix math improvements, etc. Inference should be significantly faster depending on how high they get the memory bandwidth and what they do with the interposer and SoIC / chip stacking, if it happens this generation.

There are metal optimized libraries that work pretty well and there is an apple-funded CUDA -> MLX workflow in progress.

This machine is going to be excellent for personal and professional AI development / fine-tuning, and for ~$12k it’s going to cost what one nvidia card with 1/5 of the GPU memory does.

If you need really big data nvidia stuff you can just connect to the cloud, but for local work nothing will compare to this for a long time that isn’t 10x the cost.

I say all this also owning a 5090 that’s going into a Linux workstation next week for local CUDA work by the way. :)


Now, fi you were talking about Apple Intelligence… fair point! I have low faith especially given the new rumors that they are not teaming with Anthropic which is an enormous mistake, if true.

Apple’s small models are doing pretty cool stuff but they’re going to take a couple more years to be really relevant relative to the best of what else is out there now, and that is also dependent on them retaining talent. They are doing groundbreaking R&D work with on-device models particularly in RAM constrained areas, but everywhere else they are trailing pretty badly.
 
Last edited:
Waiting for M6 2nm will be good new packaging new architecture this current architecture already same from A16
Keep in mind M6 proably won’t get an Ultra variant, if you want that level of performance you’re probably going to have to wait until late 2027 or mid 2028.

My suspicion is they are going to do an Ultra every other chip cycle given the demand and yields, and with a new process node… seems like a skip probably.

Also, M5 is based on the A19, it has the same security features exclusive to that design. We know this already based on the recent release. It’s (I think?) the first time they are at parity within weeks of each other.

edit: if you're talking about the lithography, to a degree that's true albeit there have been revisions to the 3nm process, but probably nothing that will approach 2nm which should come next year.
 
Last edited:
Nice! This is pretty much what I have been waiting for to replace my main archive and mural building machine with.
The first thing I thought of when you said "mural building"....

a mac studio as the head of a robot that physically builds the murals. :)
 
Last edited:
  • Like
Reactions: TalkieToaster
Has it been confirmed that the M5 has the UltraFusion connector?
There has been a rumor that the reason the M5 Pro and Max chip chips did not come out this fall was that they were using a new fabrication technique with them that allows for a more flexible packaging of mixed components. Current techniques require brut force methods like ultra fusion but this new packaging is supposed to allow them to add RAM and GPU without needing to duplicate the whole chip. Presumably that would apply even more to an M5 Ultra.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.