iPhone 15 Pro The iphone 15 pro is about as fast as it will ever get. 2nm or possibly 1.8nm will be the end of the line for miniaturization and Moore's law.

[73CortinaV8]

Within 2 years or so CPU processes will not get any smaller and will be at maximum transistor density. Phones may not get any faster or efficient for years or even decades until some new exotic process is developed. Maybe never.

So the A18 and the A19 will be a tiny bit faster. Probably not much. Apple might be able to pack more cores into a die but that will take up space and generate more heat.

 

[now i see it]

This prediction was made 30 years ago regarding computer CPUs

 

[Starscape]

I didn’t realize we had an electrical semiconductor engineer here. Your prediction is so wrong that it’s almost funny. Well, actually it is funny.

 

[73CortinaV8]

Moore’s law: further progress will push hard on the boundaries of physics and economics – Physics World

James McKenzie pays tribute to the foresight of Gordon Moore, the founder of Intel who died earlier this year

physicsworld.com

 

[Wizec]

From the article:

“Germany, Japan and the US made a transistor consisting of one molecule of phthalocyanine surrounded by just 12 indium atoms, which – with a gate size of 0.167 nm – would be “the absolute hard limit for Moore’s law” (Nat. Phys. 11 640).”

However, this level of miniaturization may not be economically feasible, even though its is technically possible.

 

[Wizec]

Newer technology may supplant silicon chips though:

Photonics researchers report breakthrough in miniaturizing light-based chips

University of Rochester team demonstrates new way to control light as it moves through integrated circuits.

www.rochester.edu

 

[Zest28]

3nm is not really 3nm, it is just marketing gimick to impress customers. A 3nm chip is not 3nm thin.

The next gen chips will be faster than the A17, you will see.

 

[darkus]

This prediction was made 30 years ago regarding computer CPUs

Yup. This prediction has been made so many times in the past and its always been proven wrong. I mean when people were playing with 286 and 386, who could have thought multicore intel chips would be around, let alone arm, let alone whatever we will see in the next 5 years. Have no doubt about that

 

[sack_peak]

NDAs exist to make these statement factual until they are not.

> 2nm or possibly 1.8nm will be the end of the line for miniaturization and Moore's law.