If there are a lot of travelers that do LLM training, there might be a market for it. A workstation laptop that is about 1” thick. Maybe 10k units per quarter? Per year?
They need more workstation software to really give it a go. Don’t see them seeing it worth the effort without some inroads or momentum with workstation software. Things like ML, LLM, CAD, rendering, numerical modeling, language support, etc,
I would think something like 50k units per quarter as worthwhile. That’s like 500k MBA units.
If there are a lot of travelers that do LLM training, there might be a market for it. A workstation laptop that is about 1” thick. Maybe 10k units per quarter? Per year?
They need more workstation software to really give it a go. Don’t see them seeing it worth the effort without some inroads or momentum with workstation software. Things like ML, LLM, CAD, rendering, numerical modeling, language support, etc,
I would think something like 50k units per quarter as worthwhile. That’s like 500k MBA units.
What are you workstations use cases?
I was thinking what me and my studio want. Basically a portable Mac Studio laptop for live events, on location filming with editing the same day, multitasking graphic design, after effects, and video editing, 3D modeling, and then all the basics with minimal ai. Time is of the essences d every second saved adds up.
Is there no longer a market for that?
Even if it were “only 50k units per year, that’s a pretty substantial business no?
Especially considered it they don’t have to change much. A new screen and increase chassis dimensions. Use an existing SOC and done. Next thing you know the ultra chip is selling at a better clip.
And if the screen size doesn’t change, there is hardly any effort at all. Just stronger cooling.
Seems like a super easy way at to cover the mobile po er user and sell another 50k (perhaps much more) ultra chips, which are a bit of a pain point in selling umbers currently. This would add the benefit of helping the investment case for monolithic ultra dies, allowing the SOC to function much more efficiently.
You're asking for an M4 Max. That's the portable compromise.
If you need more than the laptop form factor can accomodate, rack-mount the high-end desktop version — an Ultra Studio — and drop it in a flightcase. That's been the solution for the past twenty-five years.
You're asking for an M4 Max. That's the portable compromise.
If you need more than the laptop form factor can accomodate, rack-mount the high-end desktop version — an Ultra Studio — and drop it in a flightcase. That's been the solution for the past twenty-five years.
No. I’m asking for an m4 ultra. Says so in first post.
The max is great
but at 1.4 nm, it should be feasible for apple to get an ultra chip in the MacBook pro. Even more so if there is a bigger model.
There are many time-sensitive projects, especially increasing capabilities at live events, that can benefit from all the horsepower one can possibly throw at it. Often time. The working environment is tiny and mobile. Cutting near real time footage while red design motion graphics and customizing designs on the fly taps out just about everything. Something like an ultra in a form factor that you can move around with at an event would be HUGE.
You're asking for an M4 Max. That's the portable compromise.
If you need more than the laptop form factor can accomodate, rack-mount the high-end desktop version — an Ultra Studio — and drop it in a flightcase. That's been the solution for the past twenty-five years.
No. I’m asking for an m4 ultra. Says so in first post.
No, you're asking for a MacBook. Says so in first post.
The Max is the portable compromise. You're not going to get an Ultra in laptop form factor.
The portable Ultra production machine is a rackmounted M3 Ultra Studio in a flightcase.
You're asking for an M4 Max. That's the portable compromise.
If you need more than the laptop form factor can accomodate, rack-mount the high-end desktop version — an Ultra Studio — and drop it in a flightcase. That's been the solution for the past twenty-five years.
No. I’m asking for an m4 ultra. Says so in first post.
No, you're asking for a MacBook. Says so in first post.
The Max is the portable compromise. You're not going to get an Ultra in laptop form factor.
The portable Ultra production machine is a rackmounted M3 Ultra Studio in a flightcase.
No. I’m asking for an Ultra series MacBook. Says so in the first post. Note that such is very different from the m4 max you were pushing.
It’s a legit request for thoughts on such a setup from apple and the use cases it would be exceedingly valuable in. I imagine it would sell a little better than the ultra Mac Studio, but not groundbreaking. Yet it would add to the number of MacBook sales, which is only a good thing.
An ultra chip in a MacBook Pro would very likely be feasible at 1.4 nm. Certainly would set records.
You seem to be focused on the current 3nm process s as he m4 generation. I’m talking two generations from now.
And if there was a 18”-18” screen option, that would be great as well.
If there are a lot of travelers that do LLM training, there might be a market for it. A workstation laptop that is about 1” thick. Maybe 10k units per quarter? Per year?
They need more workstation software to really give it a go. Don’t see them seeing it worth the effort without some inroads or momentum with workstation software. Things like ML, LLM, CAD, rendering, numerical modeling, language support, etc,
I would think something like 50k units per quarter as worthwhile. That’s like 500k MBA units.
Well you have to wonder how many Mac pros or ultra series studios sell each year.
I imagine the ultra series MacBook Pro would outsell both -combined.
And that just pads tje overrall MacBook Pro lineup as the ultra does the studio, and as the pro does the Mac mini.
If there are a lot of travelers that do LLM training, there might be a market for it. A workstation laptop that is about 1” thick. Maybe 10k units per quarter? Per year?
They need more workstation software to really give it a go. Don’t see them seeing it worth the effort without some inroads or momentum with workstation software. Things like ML, LLM, CAD, rendering, numerical modeling, language support, etc,
I would think something like 50k units per quarter as worthwhile. That’s like 500k MBA units.
Well you have to wonder how many Mac pros or ultra series studios sell each year.
I imagine the ultra series MacBook Pro would outsell both -combined.
And that just pads tje overrall MacBook Pro lineup as the ultra does the studio, and as the pro does the Mac mini.
I don't think they get enough Ultra system sales as they don't update those systems on a yearly basis. We'd know really, as there would be momentum in workstation software being ported and optimized for Apple's CPU, GPU, NPU core and memory designs. People would be buying Ultra systems in good numbers. Apple would have gotten the M3 Max/Ultra Studio and Pro machines out the door last year if so.
I would agree that a Mac laptop with an Ultra SoC will get more sales than the Mac Pro and probably the Mac Studio Ultra. It's not going be a good laptop in Apple's view I think. Laptops are basically limited to 100 WHr batteries because of airline regulations. A laptop would have to remove 2x the heat (running at about 250 W or 1000 BTU/hr) of the maximum in the MBP16. So, some combination of increase flow rate, increased heatsink fins and materials to get that doubling, and likely means a 30 mm thick laptop.
So, probably 3 to 4 kg, 2x the thickness, and half the runtime. They'll get over the inelegance of it if the sales are there. Their current content creation and web worker market really can't use the performance. It really needs workstation software. If Matlab was a first class citizen on macOS, an actual macOS app, that would be a very good sign.
I do think you are confused about the 1.4nm. You have to think in terms of power consumption. An Max chip fabbed at 1.4nm will have about the same power consumption as a M4 Max and fit inside their current MBP form factors, running at about 125 W by design. This chip will have about the same performance as an Ultra chip fabbed at 3nm and perhaps even 2nm.
Apple can design an Ultra SoC using 1.4nm chips, and its max power consumption will be 250 W, which would require a laptop that is 2x the thickness, weight, and have half the runtime. They are all moving targets in terms of performance. Their power consumption is what really determines what size laptop they can fit into. Ultra chips can only fit in what are classified as workstation laptop form factors or gaming laptop form factors. 30 mm thick, 3, 4, 5 kg, less than 5 hours of battery life because they use a maximum of about 250 W.
Comments
They need more workstation software to really give it a go. Don’t see them seeing it worth the effort without some inroads or momentum with workstation software. Things like ML, LLM, CAD, rendering, numerical modeling, language support, etc,
I was thinking what me and my studio want. Basically a portable Mac Studio laptop for live events, on location filming with editing the same day, multitasking graphic design, after effects, and video editing, 3D modeling, and then all the basics with minimal ai. Time is of the essences d every second saved adds up.
but at 1.4 nm, it should be feasible for apple to get an ultra chip in the MacBook pro. Even more so if there is a bigger model.
The Max is the portable compromise. You're not going to get an Ultra in laptop form factor.
The portable Ultra production machine is a rackmounted M3 Ultra Studio in a flightcase.
I would agree that a Mac laptop with an Ultra SoC will get more sales than the Mac Pro and probably the Mac Studio Ultra. It's not going be a good laptop in Apple's view I think. Laptops are basically limited to 100 WHr batteries because of airline regulations. A laptop would have to remove 2x the heat (running at about 250 W or 1000 BTU/hr) of the maximum in the MBP16. So, some combination of increase flow rate, increased heatsink fins and materials to get that doubling, and likely means a 30 mm thick laptop.
So, probably 3 to 4 kg, 2x the thickness, and half the runtime. They'll get over the inelegance of it if the sales are there. Their current content creation and web worker market really can't use the performance. It really needs workstation software. If Matlab was a first class citizen on macOS, an actual macOS app, that would be a very good sign.
I do think you are confused about the 1.4nm. You have to think in terms of power consumption. An Max chip fabbed at 1.4nm will have about the same power consumption as a M4 Max and fit inside their current MBP form factors, running at about 125 W by design. This chip will have about the same performance as an Ultra chip fabbed at 3nm and perhaps even 2nm.
Apple can design an Ultra SoC using 1.4nm chips, and its max power consumption will be 250 W, which would require a laptop that is 2x the thickness, weight, and have half the runtime. They are all moving targets in terms of performance. Their power consumption is what really determines what size laptop they can fit into. Ultra chips can only fit in what are classified as workstation laptop form factors or gaming laptop form factors. 30 mm thick, 3, 4, 5 kg, less than 5 hours of battery life because they use a maximum of about 250 W.