Looks increasingly likely that a M5 device will make an appearance at WWDC. With MacBook Air launching in the next few weeks with M4, it is possible to see an iPad Pro with M5 at WWDC. Initial volumes required for an iPad Pro will be less too.
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.
M5 Apple Silicon Chip Reportedly Enters Mass Production
- Thread starter MacRumors
- Start date
- Sort by reaction score
If you do heavy machine learning training, the power is never enough. True that, if you need *a lot* of power for ML, you'd probably better rent some cloud resources.
LLMs are asymmetric. Local inferencing is possible now with decent sized models that are "open source". Open sourcing made this possible, btw OpenAI never was open. On the other hand, LLM training at a decent scale was never possible and will never be. Cost of Nvidia H800 is estimated at $2/hr and you need millions of hours.
m5 base ipad pro = october of 2025.
m5 pro/max mac book pro , end 2025 makes more sense
I'm ready for an M5 Pro MBP. I have an M1 Pro MBP now and it's finally starting to show it's age a bit in my normal workload.
!
Destroying the Mac? How? In the most recent quarter Mac revenue was up 15% at $9b. Maybe the Pros you describe are finding they can do their work with one of the current lineup.
Not abandoned. Not sure who thought Apple would upgrade the Ultra variant as often as everything else. Due to low sales volumes, it’s a very costly SoC to produce, so expecting a turn around like the other SoC variants isn’t realistic.
Furthermore, the M2 Ultra Studio is hardly outdated it still sits at/near the top of performance levels.
I explained it….ignoring pro users has a cascade effect. I literally stated this in the post.
I think this is probably correct. An Ultra cost a lot of resources to produce relative to the low sales, so to amortize the development, they have to keep people buying it for longer until development costs are recovered and that justifies the next version. Doesn't have to be a cash cow like an iPhone, but don't dig a financial hole with the Ultras.
The only problem with this scenario is that Studio sales have probably mostly dried up. And Studio owners, or people who need Studio, or multiple Studios for compute, are withholding their dollars and just waiting.
News flash: next year will be 2026
Which is why the article makes little sense at the bit I quoted.
Gaming.
Running Windows Arm version under a virtual machine with fairly demanding ap.s.
Same here. My M1 Pro still doesn’t feel slow in anyway so I may hold off until the actual laptop design gets a refresh.
I’m more looking forward to a TB5 Pro Display XDR with Pro Motion (120hz+) than a TB5 Studio Display.
The Studio Display the most mediocre prosumer product they sell—if it can even be called a prosumer product not having HDR support majority of key creative professionals want to produce and create their work that Apple supports.
It’s incompatible and not on par with the rendering capabilities of any of Apple’s prosumer products with a screen:
The Studio Display’s screen is inadequate to render correctly all the HDR content that the iPhone Pro, iPad Pro, Macbook Pro, and the Vision Pro can.
It doesn’t have 1000 sustained nits nor 1600 peak nits to be on par with your use of any of such devices either making it an indefinite mismatch using such devices and expect consistent screen performance and output.
Last edited:
I bet Apple didn't have a 5K OLED panel available at that time. Maybe one is available now. If so, I bet the price increase will be beyond reach for many people. If there is a big price increase, well, you already have the 6K Apple Display. The Apple Studio Display seems to sell well. I know plenty of people who have one or two. (I don't have one. I have a Samsung optimized for screen area.)
Last edited:
How strong is the market demand for such a product? It seems just lately there are a couple more 6K 32" displays on the horizon; to get Thunderbolt 5 AND 120-Hz refresh rate sounds like a long shot expectation, and the XDR is already very expensive (at least from the mainstream consumer standpoint).
I'm a consumer and don't produce content, so I'm curious about this. When I was monitor shopping late last year, reviews of IPS panel-based displays downplayed HDR as basically something that had mild effect on those displays as a class.
I was under the impression if you really want HDR, you go VA panel (lousy viewing angle breadth) or OLED.
Is there an IPS-based display you think is strong on HDR of a brand the mainstream public is likely familiar with?
Ironically the ASD's 600 nits is brighter than many competing displays in terms of max. brightness.
For people with the use case scenario you have in mind, what are they using for displays today?
MiniLED is the alternate technology to OLED even competitors as well as Apple used on the current Pro Display XDR to even offer the same peak nits AND sustained nits along with more open and standardized means of tweaked the color calibration—built in colorimeter such as Asus Pro Art series.
A large OLED panel equivalent would indeed will be expensive; the Studio doesn’t have to use such tech.
The cost of matching Apple’s Pro Display XDR HDR specs after 5 years now is a bit exaggerated.
At minimum gaming monitors at $1200 now even have Dolby Vision and much, much higher peak nits and sustained nits (underrated) than the studio display.
That said, it is not trivial or easy to quantify the costs of 5K and 6K to manufacture.
It’s much cheaper now than before as one would expect when the Pro Display XDR and Asus PA32UCG (Asus’s more versatile equivalent but at 4K@120hz with VRR to accommodate game devs and animators) over 5 years ago; both had a MSRP of $5000.
DisplayPort 2.1 and HDMI 2.1 enables both their successors to have 120hz + Dolby Vision HDR support.
HDR is overall very important; most obvious impact of picture quality besides PPI.
As far as demand: There is absolutely demand that bottlenecks people embracing high PPI.
People want monitors with no functional compromises which necessitates high PPI and high refresh mates able to accommodate a wider variety of prosumers (animators, game devs, other kind of devs, digital artists, content creators, an etc that may also game) and wider high-end demographics such as gamers.
High PPI monitor adoption is bottlenecked by the lack of high refresh variants to be much more appealing as a whole.
Last edited:
That's true. Even with a top end M4 Max with 128G RAM, the shader benchmark is not smooth at all.
I also have a Dell workstation with a dedicated RTX A5000 GPU with 24G vram. It's way smoother. But a dowside of the GPU is the limited vram. 24G is not small, but for AI workload, such as running LLM models locally, it's still a limiting factor. I can only run 32b models, such as Gemma2 32b, anything larger wouldn't run on GPU and CPU is just too slow for that.
The unified memory architecture in Apple Sillicon shines here. With the 128G RAM, I could at least run llama3.3 or deepseek-r1 with 70b parameters and it's at least usable. On the Dell, I couldn't even load such models.
…How it isn’t to you considering the make up of the panel and the amount of zones aligning with it being so?
The Pro Display XDR is a FALD display, not mini-LED, as there are an insufficient number of zones to make it mini-LED. All mini-LED is FALD, but not all FALD is mini-LED.
The number of zones in the Pro Display XDR is measured in the hundreds, but it is measured in the thousands for the MacBook Pro, despite the screen of the MacBook Pro being many times smaller.
…Mini-LED IS by definition known to be a Full Array Local Dimming (FALD) panel though.
The amount of zones is an implementation detail that makes them distinct than traditional panels.
A smaller and newer mini-led implementations increasingly have more zones than before—especially smaller ones that will achieve that sooner and will be produced more by panel manufacturers
Is it a similar situation where monitor pundits believe “tandem OLED” lost its meaning with far more advanced tandem implements than before?
The introduction of mini-LED launched displays like the TCL 8-series which had tens of thousands of LEDs in about 1000 zones, and this thousand(s) of dimming zones became one of the defining features of mini-LED.
In contrast, high end FALD TVs of the same time frame were in the hundreds of dimming zones. The Pro Display XDR, which came out around the same time frame as the first mainstream retail mini-LED TVs, had far fewer dimming zones than those mini-LED TVs, again measuring in the hundreds like other FALDs. To be fair, the Pro Display XDR was never marketed as a mini-LED display, and press did not refer to the Pro Display XDR as a mini-LED display either.
So you may think this is an arbitrary distinction and you would be right about that, but again based on numbers of dimming zones, all mini-LED is FALD but not all FALD is mini-LED. The Pro Display XDR falls into FALD category but misses out on the mini-LED subcategory.
I have no idea where you're going with this. Tandem OLED means 2 or more OLEDs stacked. In Apple products this only exists in the iPad Pro. Better tandem OLED in the future will also be tandem OLED, but better.
OTOH, yes the Pro Display XDR has more dimming zones than early FALD displays, but does not have as many dimming zones as the early mini-LED displays. Thus, it has never been categorized in the mini-LED category.
Last edited: