When to Expect New M4 MacBook Air, Mac Studio, and Mac Pro Models

[sw1tcher]

No, it means that Apple didn't expect AI to become so popular so quickly, so they had to rush developing the M4 and market it as "the best chip for AI ever."

It’s not like Apple to miss out on these marketing opportunities.

I thought that was the M3 chip?

Apple unveils the new 13- and 15-inch MacBook Air with the powerful M3 chip

Apple today announced the new MacBook Air with the powerful M3 chip, taking its power-efficient performance and portability to a new level.

www.apple.com

Apple today announced the new MacBook Air with the powerful M3 chip, taking its incredible combination of power-efficient performance and portability to a new level. With M3, MacBook Air is up to 60 percent faster than the model with the M1 chip and up to 13x faster than the fastest Intel-based MacBook Air.1 And with a faster and more efficient Neural Engine in M3, MacBook Air continues to be the world’s best consumer laptop for AI.

With the transition to Apple silicon, every Mac is a great platform for AI.

 

[MacLawyer]

This seems to confirm the M3 chip was basically a complete failure across the board and Apple is moving away from it as fast as they can.

I kind of like my M3 MBA. 😄

 

[MacLawyer]

The new MacBook Air really needs to meet or exceed the Surface Pro laptop at the same price point. And should really have been released before Christmas/Holidays this year.
View attachment 2404305

At least they put a decent amount of RAM inside.

 

[andrewxgx]

any idea whether desktop M4's will have Thunderbolt 5? this would open doors for 120 Hz 5k Studio Display.

 

[sw1tcher]

Apple will update its MacBook Pro, Mac mini, and iMac lines with its latest M4 chip as early as this year, according to Bloomberg reporter Mark Gurman, with the Mac mini also set to feature its first redesign since 2010.

Is "as early as" the new "but obviously things can still theoretically change"?

"As early as this year" doesn't mean the MacBook Pro, Mac mini, and iMac will for sure get an M4 update this year. It means it's possible they could be updated this year, but it's also possible they won't be updated this year.

It's similar to when a car dealer advertises a financing rate "as low as" 1.9% APR. It doesn't mean everyone will get a 1.9% APR.

 

[Pakaku]

No defense for Apple, they should have been ready for a trillion dollar company.

Money can’t buy everything

 

[ILoveCalvinCool]

I don’t have experience with Mac myself, but adaptor galore is Apple’s style, no? MBA also only has USB-C, right? So why would a Mac Mini require HDMI, power and Ethernet connections? That all fits into 1 USB-C port. So make it 1-2-3 USB-C ports and they’re done, I’d say.

If Apple starts stripping ports from their meager line of desktops in order to make them smaller and thinner, I will personally call upon the hounds of hell to torment Tim Cook until the end of his days.

 

[ajmas42]

I find it odd when people say Apple missed the AI bandwagon, when their chips have sported the Neural Engine for a while. I’d argue they were ahead of the curve, but because they weren’t using the letters AI, people assumed Apple hadn’t ventured in there.

When it comes to the new chips, what I care about most is improved single threaded performance and better GPU capability. If it can play a PC game at what feels like native performance, then for me that already a good measure is success.

 

[Zdigital2015]

No defense for Apple, they should have been ready for a trillion dollar company.

A bazillion dollar company cannot change the laws of physics or force something to be ready before it’s ready…this isn’t an episode of Star Trek where Scotty or Geordi comes up with a last minute fix for a warp core breach. Try to live in the real world.

 

[satchmo]

While I'm glad Apple isn't stagnant and continually improving chip silicon, I just wish it was simplified.

So you've got M2, M3, and soon M4 all available in Base, Pro, Max and Ultra, and possibly Extreme variants.
So five options within each chipset.

As someone who is still on a M1 Mac mini, I find the number of options to upgrade dizzying and paralyzing.

 

[ksec]

Your tech knowledge far exceeds mine, but from a business, planning and strategy perspective, M3 was not a success.

Well, the short life span of M3 is precisely because without M3 there will never be M4. So from a business, planning and strategy perspective. It is perfect.

 

[genzai]

This seems to confirm the M3 chip was basically a complete failure across the board and Apple is moving away from it as fast as they can.

you are hilarious- this confirms what was always widely reported, that N3B was a first gen, relatively expensive process and that apple and there rest of the industry is moving to N3E as quickly as possible. M3 is a fantastic ship- but moe that they have M4 ready and shipping there is obviously no reason to release new Macs with M3 and the older process.
g\

 

[FelixDerKater]

iPad should not be the first to get the newest processor line.

 

[anthonylambert]

Apple have a huge problem which is not easy to solve. The market is pushing them every year to produce better chips in the whole range, phones and macs.

The simpler chips are developed first and so are ready to be in products like the air and low end Mac books.
then it takes another 12 months to get the higher end Mx chips in the higher end hardware.

The new generation chips at the low end are better than some of the medium end chips of the old generation. So they are canibalising the higher end machines. in some cases for over a year. Mac studio?

They are trying to move all processors to a yearly release.

They should wait until all the chips are fully developed and release the machines high to low.

They would then sell more high end machines as people may buy the more expensive machines as they can't wait for the next Air or Mac Book, Mini.

All this is tempered by the market shares of the machines. Meaning currently Apple makes its money from Airs and Mac Books / MB Pros.

 

[anthonylambert]

I'd like Apple to produce Mx processor and memory boards for the Mac Pro.....

You buy yr Mac Pro then if you want to have more processor you just add a PCIe board which can take maybe 4 added daughter boards with an Mx and memory. This would be great for AI machines. Or HPC. Apple could afford to sell these addon Mx chips fairly cheap (HaHaHa - dream on) and still make lots of money.

 

[Zdigital2015]

Your tech knowledge far exceeds mine, but from a business, planning and strategy perspective, M3 was not a success.

The M3 was a stopgap measure because otherwise there would be complaining on these boards that Apple was taking to long to update the M2. It wasn’t a failure in my mind as much as an interim step that still has its merits, but after the introduction of the M4, which shocked me a bit as well as a lot of others, it definitely became apparent that M3 was not long for this world.

I completely understand why Apple opted not to make the M3 Ultra and I’m sort of surprised they even made the M3 Pro and the M3 Max other than both of those show how Apple is evolving their design philosophy for those SoC’s for better or worse.

M3 is the end of the first gen of M-Series, in my opinion. Better days are ahead for sure, the future is bright.

 

[Corefile]

Not trying to be a jerk. But. How exactly was the M3 chip a failure?

It might be viewed as analogous to that 10nm monstrosity that Intel released which was quickly EOL'd. Basically a marketing release to bridge the gap and save face.

 

[ScholarsInk]

Why do this? Among other things, the Macs touted as "most powerful Macs ever" could actually hold that crown for the entire generational cycle instead of only 3 or 4 months at most.

I was definitely pissed when I bought a fully loaded Mac for the first time (M2 Max) only for M3 to launch months later. It’s not even a “most powerful” ego thing; I just felt ripped off.

 

[Kay_Ess]

I find it odd when people say Apple missed the AI bandwagon, when their chips have sported the Neural Engine for a while. I’d argue they were ahead of the curve, but because they weren’t using the letters AI, people assumed Apple hadn’t ventured in there.

When it comes to the new chips, what I care about most is improved single threaded performance and better GPU capability. If it can play a PC game at what feels like native performance, then for me that already a good measure is success.

That’s comparing apples to oranges. What Apple does with the Neural Engines is something completely different than what ‘we’ consider AI nowadays (generative pre-trained transformers) The NE for example can tag your photos so you can search for pics of your cat by typing in his name or you can ask Siri to set a timer and you can just use efficient and tiny chips for that. If you want to run a small GPT however you’ll need a really serious processing chip accompanied by 32-64GB of RAM. That GPT can be used to converse with, for text and image generation, translating, text completion, question answering, sentiment analysis, code generation, and more.

They’re completely different animals even though it’s all grouped under ‘AI’

 

[megastuff999]

You may as well ask a fortune teller as listen to anything Gurman spouts.

 

[skitidetdu]

Very curious about the Mac mini m4 re-design but mostly how the m4 pro will perform with Logic Pro since it relies only only performance cores. Is it known what the number of efficiency/performance cores the pro chip will have?

 

[HobeSoundDarryl]

I was definitely pissed when I bought a fully loaded Mac for the first time (M2 Max) only for M3 to launch months later. It’s not even a “most powerful” ego thing; I just felt ripped off.

Now we think we know a rough schedule of when new Macs will hit... but it's actually changed in every generation so far, so we really know very little with any confidence... as is always the case when people think the see pattens in only a few iterations. I can flip a coin a few times and it may come up heads each time. Who wants to bet the farm the next flip will also come up heads?

However, I very much believe that what makes the most sense of all angles is actually flipping the launch schedule as described... basically leaving the "most popular" Macs in that Fall release of roughly OCT but putting the most powerful M-chip Macs launches AHEAD and essentially working down to cheapest Macs. Thus,

M-Ultra (and hypothetical Extreme) FIRST (maybe Spring-to-WWDC)
M-PRO & M-MAX FALL (usually about OCT)
M-base (either)

Example:

• M5 Ultra/Extreme next WWDC
• M5 PRO & MAX next October
• M5 the Spring after that

From Apple's (seller) perspective:

• Most profitable Mac (even if volume is light) rolled out FIRST, so they are king of the power hill for the whole cycle and those who desperately need most powerful can't wait a few months to get it for a lot less money.
• Most popular Macs are still launched at the same time
• Least profitable Macs are rolled out last (against the backdrop of more powerful, next gen Macs right around the corner)

As is, I'd never buy another Ultra without such a change, knowing that the next gen MAX is only a few months away. Instead, I'd just wait for that MAX, spend far less money and scratch my "most powerful" itches. Flip the schedule as described though and then Ultra (and Extreme) make much more sense... perhaps even pulling some of those "wait for MAX" people up to paying much more for an Ultra or Extreme and thus making the Mac they buy much more profitable for Apple. Few if any are going to buy BOTH, so this seems to be the way to maximize profits on all Mac releases... though it's not been Apple's schedule so far.

If this rumor is true, M4 Ultra at WWDC, then M5 MAX in October??? Even for the monster hop from M2 to M4, why bother? Just wait for M5 MAX only a few months later.

 

[bradman83]

2) With AI dominating the headlines almost every single day, it simply makes sense to rush to an "AI-capable" chip. Sure, the M3 variants can likely do SOME AI, but the M4 has the supposedly-necessary bump to do it better.

Minor quibble - the neural engine on the M4 in real world usage is only about 5% faster than the neural engine on the M3.

The problem with "TOPS" or trillions of operations per second ratings is that the number depends on the operation in question. The M3 was rated with FP16 operations whereas the M4 switched to INT8. The two aren't directly comparable but, on average, an FP16 operation can do twice the work of an INT8 operation. Again not entirely comparable, but that's why the M3's neural engine is rated at 18 TOPS (at FP16) versus the M4's 38 TOPS (at INT8), whereas in reality the performance boost isn't nearly as big as the advertised numbers would have you believe.

 

[ryanmp]

.... and the new larger iMac (not the kiddy play school version). Come on Apple, it's time.

 

[deputy_doofy]

Minor quibble - the neural engine on the M4 in real world usage is only about 5% faster than the neural engine on the M3.

The problem with "TOPS" or trillions of operations per second ratings is that the number depends on the operation in question. The M3 was rated with FP16 operations whereas the M4 switched to INT8. The two aren't directly comparable but, on average, an FP16 operation can do twice the work of an INT8 operation. Again not entirely comparable, but that's why the M3's neural engine is rated at 18 TOPS (at FP16) versus the M4's 38 TOPS (at INT8), whereas in reality the performance boost isn't nearly as big as the advertised numbers would have you believe.

I'm good with that minor quibble.