Jury's out for M4 Ultra chips on Apple Watch 11 though.
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.
Here's Why Apple is Unlikely to Release an M4 Ultra Chip for Macs
- Thread starter MacRumors
- Start date
- Sort by reaction score
People said that's not true, but anyway, if I remember correctly they also promised an M1 Mac Pro that never happened.
In the car industry and others, it's pretty normal to promise something and then just change plans because it's too expensive. Or even to say you'll make something just to see how potential buyers react.
Apple is more solid than some rivals but that's mostly because of iPhones. Macs are a side-business that can act a little differently.
Couldn't they just eliminate the Pro Chip and replace it with the Max (with UF) to increase scale and reduce cost if that's the issue? Apple's regular CPU's are good enough for the vast majority. While 3D rendering, 4K video editing, scientific simulations, LLM and data analysis professionals usually opt for the Max CPU as a minimum from my observations but I could be wrong.
This right here - the pro market used to need those high end chips. That changed with Apple Silicon. the Max series are so absurdly powerful that the average professional photographer/videographer can use an M1/2/3/4 Max machine with 32+ GB of RAM and get massive improvements from the Intel chips.
Those of us who use our computers to predominatenly write papers, organize photos in the Photos app, send emails, etc... don't need a Pro or a Max, and can get by just fine on an M4 machine (air/pro/mini). Even the average podcaster could get a very powerful M4 or M4 pro machine and do what they need to do.
Apple silicon has brought "pro-level" chips to the mass market with the M4 chips, the Pro and Max versions are insanely fast and powerful, the Ultra just isn't needed.
My 2¢
Why let people buy an M4 when you can charge them more for an M4 Pro that they'll never take advantage of, but it's a "Pro" so people feel like they want it.
There are so many posts about the M4 Ultra lately. Pretty funny because absolutely nobody here needs that kind of performance.
True. If they really did something logical that wasn't driven by profit, they would now discontinue the Studio as the new Mini has proven to be a powerhouse in a nicer form factor, and save the Ultra exclusively for the Mac Pro to justify its price tag.
MBA - M standard
MBP - M standard and upgradeable to Max
Mini - M standard and upgradeable to Max
Studio - discontinued
Mac Pro - Ultra
But I also don't know anything about running a company with a 5 trillion dollar market cap.
It's possible the M3 Max chips on the M3 MBP's do indeed lack the UltraFusion bridge, and that the M3 Max chips used in the Ultra are simply a different variant—as suggested by the fact that the M3 Max MBP's are TB4, while the M3 Ultra is TB5.
That doesn't necessarily mean they updated the on-die TB controllers on the M3 Max for the chips used in the M3 Ultra (there may be ways to upgrade the M3 Max chip to TB4 without updating the on-die TB controllers), but does strongly suggest it.
Surprised that sales of the Mac Studio are "small." I would have thought many people are buying both the Max and Ultra variants. The Studio is a fantastic machine, even if your use case is AI.
1. M3 Max was also known not to have ultra fusion as well and yet they changed it which is quite weird.
2. Not a great excuse especially since it's just two Max chips.
3. SoC is the biggest problem since the die size is just too big. A big die means too expensive and difficult to manufacture. This is why chips are aim to make smaller. Perhaps Apple realize now that SoC is not a good solution and maybe should go for chiplet or McM.
2. Not a great excuse especially since it's just two Max chips.
3. SoC is the biggest problem since the die size is just too big. A big die means too expensive and difficult to manufacture. This is why chips are aim to make smaller. Perhaps Apple realize now that SoC is not a good solution and maybe should go for chiplet or McM.
For that matter, an M1 will do all that routine stuff. Apple could do a die shrink of the M1 to 3 nm and go for even more battery life on a "low end" laptop, but they don't want that market.
The Linux desktop I use for my main computer is roughly M1 speed. It's also fast enough, I'm not CPU bound. It does have 16 GB RAM (with two open slots for more), 3 internal drives(with room for two more), and 11 USB ports.
The Mac Pro has few reasons to exist, as I see it.
On tower workstations, the space is used for powerful GPUs (multiple) and lots and lots of memory - you can get them with terabytes of memory(!). Neither of that is possible on the Mac Pro.
You can add some special connectivity cards if Thunderbolt 5 and 10 GbE is insufficient, but that's a tiny fraction of a tiny segment.
Agreed. I have the 27” iMac with an i9 CPU. I just ordered a Mac Studio with M4 Max. That will last for years to come.
I wonder who qualifies as a pro user anymore. I’m not criticizing the idea. It’s just that a Mac Studio has a huge amount of computing power. It blows away the older Mac Pro with G4 or Intel CPU.
I work with large arrays of x-ray spectra. The Studio will be awesome. Others working in geology or meteorology or other should also be happy. At what point do you step up to a small supercomputer instead of a desktop?
I always wondered if every Max chip started life as part of a fully formed Ultra die and the UltraFusion connector was just part of the lithography. If both sides of the package have matching capabilities/binning and the connector is good it can be an Ultra chip. Otherwise they cut them at the connector and one or both packages become a standalone Max. That would make sense then that they it takes until production matures to release an Ultra SoC.
The best match for productivity and hobby use would probably be one of the new Mac Minis with a third party screen. I've got a 40 inch ultrawide screen (Thinkvision P40w-20), and for my use it's better than anything Apple has made. As a bonus, I can just plug in my customer provided laptop when I need more than just Office 365 there.
Last edited:
Or just drop the Pro. The usual benefits of a tower workstation don't apply to the Pro - massive amounts of memory chips, multiple large GPUs.
Similar for software development.
20+ years ago, I used high end, multi-CPU workstations with RAID as a developer to keep the build process as fast as possible. Eventually, the computers got fast enough that almost any model was OK and laptops took over.
These days, unless you're into AI almost any computer will do for software development and the choice is more a consideration of screen size and quality vs weight. Just add enough memory.
This doesn't mean developers aren't pros...
Six long years, with no reduction in price, just a slight spec bump...
A more appropriate analogy for swapping CPUs would be an engine swap...
From my understanding the reticle used in manufacturing of these chips is too small for a full Ultra SoC, hence the UltraFusion connector to allow stitching two Mn Max SoCs together...
My thought is that some (with the M4 Max Apple seems to have made the decision to not add the UltraFusion connector) Mn Max SoCs are manufactured with the UltraFusion connector, and those who pass testing can go on to become Ultra SoCs, those that fail are earmarked for the Mn Max MacBook Pro and Mac Studio products...
Not sure how you arrived at that conclusion. I’m pretty sure that SoC is the whole reason why Apple Silicon performs so well. Do you mean, just for the extreme high end?
This isn’t really a fault of architecture or manufacturing difficulty, it’s a product of having a a great design that eventually outmodes the initial plan. They were expecting to need all these tiers of performance, such that they make space for the Ultra architecture, but didn’t foresee how well-suited “max” and even “pro” chips would be for a vast majority of workloads in the pro segment.
So now, any delay in the release of new variants that fit the original plan presents as an engineering problem.
I’d be pretty willing to bet that sales of the “studio” are low, because there just isn’t that high demand for the performance.
Last edited:
i think maybe something we’re forgetting about here is that engineering resources have been diverted to server chips in the wake of the AI boom.
They’re not going to divert from the flagship line (which is really only the pro and max), so the low-sales specialty tier (Ultra) is what suffers.
“Not every generation will have an Ultra” then reads a bit like revisionist commentary on product strategy. AKA, it’s a convenient explanation that makes what is happening seem like it was the plan all along.
I’d bet that, under the hood of top-level marketing and strategy goings-on, in the wake of making a killer product that redefined the boundaries of necessity, they’re trying to adjust what all the performance breakpoints should be, as well as the corresponding naming.
So what we get in the M3 Ultra Studio is just something easy based on an “older” architecture to satisfy that high-end specialty segment. At the end of the day, an M3 Ulta is still going to belch out those exports faster than an M4 max, and run large language models that an M4 Max can’t.
They’re not going to divert from the flagship line (which is really only the pro and max), so the low-sales specialty tier (Ultra) is what suffers.
“Not every generation will have an Ultra” then reads a bit like revisionist commentary on product strategy. AKA, it’s a convenient explanation that makes what is happening seem like it was the plan all along.
I’d bet that, under the hood of top-level marketing and strategy goings-on, in the wake of making a killer product that redefined the boundaries of necessity, they’re trying to adjust what all the performance breakpoints should be, as well as the corresponding naming.
So what we get in the M3 Ultra Studio is just something easy based on an “older” architecture to satisfy that high-end specialty segment. At the end of the day, an M3 Ulta is still going to belch out those exports faster than an M4 max, and run large language models that an M4 Max can’t.
Last edited:
Same was said of the M3 chip. It took awhile but it did get here
The new Mac Studio is available with mismatched M4 Max and M3 Ultra chip options. Will there be an M4 Ultra chip in the future, though? It looks unlikely.
Here are three reasons why Apple might never release an M4 Ultra chip.
First, among Apple's line of Mac chips, the highest-end Ultra chips are effectively two Max chips fused together with a technology called UltraFusion. So, the M1 Ultra chip is two M1 Max chips combined, and the M2 Ultra chip is two M2 Max chips combined. However, Apple confirmed that the M4 Max chip lacks an UltraFusion connector, so that means Apple cannot simply release a doubled-up M4 Ultra chip this time around.
Second, Apple told several journalists and YouTubers that not every generation of Mac chips will have an "Ultra" chip. The timing of Apple revealing this information suggests that we might never get an M4 Ultra chip.
The third reason was shared by Bloomberg's Mark Gurman today.
In his Power On newsletter, he said that Apple is reluctant to develop an M4 Ultra chip from scratch due to production challenges, costs, and the relatively small sales volume of its desktop computers, like the Mac Studio. So, that seems to rule out the only other way in which Apple could have released an M4 Ultra chip.
Maybe we will see UltraFusion return on the M5 Max chip, paving the way for an M5 Ultra chip?
Article Link: Here's Why Apple is Unlikely to Release an M4 Ultra Chip for Macs
How long until they hit a point without any more meaningful progress, if they really release a new generation every one or two years? Will the M11 really be much faster than the M10? What else can they to except making chips bigger and denser?