Apple Silicon's Next Big Upgrade Could Arrive in 2026

[Rychiar]

I just want m3 nowwwww

 

[JitteryJimmy]

well many here had high expectations of the 3nm. But seeing the numbers from Apple now for the A17 I don’t know what to expect for a 2nm. I am not hyped

The big difference between the gains now and the gains of 10-20 years ago is power consumption vs performance.

A17 is clearly faster and more capable, but the added power requirement is negative. That means A17 performance is at least 200% of an iPhone 11 Pro, but in a smaller physical package and using less power.

 

[iamgalt]

From the wonderlust event, even Apple themselves didn't wheel out the big wigs to announce the A17 pro chip.

I don't know how everyone else felt?

But to me the A17 pro announcement felt rather flat as if Apple themselves weren't quite as excited about this chip as we all expected them to be.

Apple's big wigs are mostly white males. Apple is more concerned about showing how diverse they are to prevent from being canceled.

 

[sack_peak]

so 3-4 years for 3-2nm, and the move from 5-3nm gave a real world increase of 0-10%. Basically looks like we'll get 10-20% increase after the next 5 years, i.e 2-3% per year?

This is why I prefer to replace my Macs, iPads & Watch 8-10 years later after their final Security Update.

I might do the same with the iPhone by 2027.

Raw performance, performance per watt and power consumption improvements would be the most pronounce by then.

At end of life hand it down or sell it to a collector.

 

[Populus]

well many here had high expectations of the 3nm. But seeing the numbers from Apple now for the A17 I don’t know what to expect for a 2nm. I am not hyped

Yeah, I was one of those, and I have to admit I’m still hyped for the upcoming N3E and especially for this new 2nm process coming on 2026. I really hope the switch to 2nm with the new transistor architecture will provide a big jump, that Apple will probably dose over several incremental upgrades… Yeah, we shouldn’t be so hyped.

 

[iBluetooth]

well many here had high expectations of the 3nm. But seeing the numbers from Apple now for the A17 I don’t know what to expect for a 2nm. I am not hyped

The 2nm will be All Around Gate, so we can expect a bigger performance improvement. TSMC didn't go for it with 3nm as the stride was too big a step. But, let's wait for actual tests on the processor as Apple may have decided to limit the speed improvement, by giving the power reduction more emphasis.

 

[ShiftHappens]

If Intel executes on time, its 18A node will already have entered mass production. But that's a big IF.

Actually looks like Intel is ahead of schedule:
- 3 nm: Sierra Forest and Granite Rapids in H2 2023
- 20 Å (2 nm): Arrow lake in H1 2024
- 18 Å (1.8 nm): Clearwater Forest and Future Lake (not sure if this is going to be the real name) in H2 2024

Since the Si atom has a 1.46 Å radius, I'd guess (like @headlessmike mentioned) we're getting close to the point where focus will shift to other areas...

 

[73CortinaV8]

moore's law is definitely slowing done.

the good thing is software hasn't caught up with hardware yet.

I wonder if Intel will ever catch up after they stumbled

 

[ThunderSkunk]

Serious question. What happens when they get to 1nm? What’s the next step? .5, or some completely new technology all together?

None of these 3nm or 2nm designations are actual measurements anymore, just branding, so they might as well call chips light years or shades of gray after they run out of nm's.

I feel like we just had this thread a week ago.
Oh yeah we did!

Perhaps they’ll call it .8nm, .5nm. …or 5/16nm, 7/64ths, 11/128s for the American market, …& then 8k picoometer, 4k femtometer, 1 quectometer…. or by then we’ll all be using quantum computing occurring on lattices of light suspended in non-newtonian fluids to check our email & watch boobs. Or sticks and small piles of sand. When the future bird archeologists 113million years from now dig up the middle primate industrial age they can decide if its worth rebuilding all the tech they’d need to decrypt our vast archives of boobies, and through some miscommunication, they’ll do it, thinking they’re finally going to understand their long lost blue-footed ancestors. And finally the culmination of our entire species intelligence into a grand cosmic joke will be complete.

 

[cocky jeremy]

Just give us bigger batteries. The performance is fine. Make it more efficient and give us bigger batteries.

 

[mectojic]

Phones get bigger, transistors get smaller...

 

[IIGS User]

Upon its creation, iOS 20 began to learn at a geometric rate. The software originally went online in June of 2025. Human decisions were removed from memoji choices. iOS then became self-aware at 10:AM Pacific time on the second Monday of September, 2025

 

[coachgq]

View attachment 2268855

Source:

- https://www.imec-int.com/en/articles/smaller-better-faster-imec-presents-chip-scaling-roadmap
- https://www.tomshardware.com/news/imec-reveals-sub-1nm-transistor-roadmap-3d-stacked-cmos-20-plans

Based on 3nm (N3) deliverables assume the die shrink to apply to the iPhone chip a year later.

Very good. Thank you!

 

[ddhhddhh2]

Alright, so it's only 2023 and they're already hyping us up for 2026? Talk about a future feature presentation!"

 

[sack_peak]

Very good. Thank you!

If you want to learn more in depth go to the links below and read the best comments.

- 2nm
- 3nm

 

[AndiG]

Useless unless Apple introduces fundamental changes in its chip design. But all the chip designers are running away from Apple ....

 

[JippaLippa]

Not even the time to enjoy the 3nm 🤣

 

[darngooddesign]

From the wonderlust event, even Apple themselves didn't wheel out the big wigs to announce the A17 pro chip.

They will bring them out with the M3s.

I don't know how everyone else felt?

Doesn’t bother me. People complain when the speeches are too long or too short.

 

[nickf]

Angstroms are next. 1 Å = 0.1 nm

Picometers if you want to keep it to SI units (i.e. 1000 pm = 1 nm). But, as others have mentioned, atoms start getting pretty large on the pm scale (the atomic radius of a silicon atom is around 140 pm), and also electrons start tunnelling in unexpected ways.

 

[darngooddesign]

Apple's big wigs are mostly white males. Apple is more concerned about showing how diverse they are to prevent from being canceled.

Yeah, it sure sucks that Apple lets everyone participate in the presentations. /s

 

[loby]

Not even the time to enjoy the 3nm 🤣

Sounds like an article to hipe down the 3nm. Everyone was looking at great jumps in performance etc. with the 3nm, but now talk begins with the 2nm, so it waters down the 3nm....

Then comes the 1nm talk so the 2nm is just a stop-gate to get to the 1nm...

never ends...

 

[Ethosik]

From the wonderlust event, even Apple themselves didn't wheel out the big wigs to announce the A17 pro chip.

I don't know how everyone else felt?

But to me the A17 pro announcement felt rather flat as if Apple themselves weren't quite as excited about this chip as we all expected them to be.

Hopefully they are putting more focus with the big wigs with Macs. M1 and M2 are decent enough (I have five including some M1 Ultra and M2 Ultra Mac Studios). But they really need to up their game on the desktop side. It’s laughable that the 2023 Mac Pro has basically 2x mobile chips

 

[mrr]

The iPhone Pro Max really needs to support the Apple Pencil.

 

[Hopefully Smarter]

Simply
don't
care as it will have no material effect on the system.

Kuo's always full of crap, yet, he's a god to you.

Examine your sources.

 

[Argon_]

The 2nm will be All Around Gate, so we can expect a bigger performance improvement. TSMC didn't go for it with 3nm as the stride was too big a step. But, let's wait for actual tests on the processor as Apple may have decided to limit the speed improvement, by giving the power reduction more emphasis.

The biggest gain from GAAFET might be the reduction in leakage. More power to computing and less to heat. We'd have even cooler phones and laptops, with longer battery life.