Image source: Apple.
Ever since Apple (AAPL 0.17%) unleashed its A7 chip onto the world, a chip that the company's marketing team claimed to have "desktop class" performance, there has been rampant speculation that Apple may someday opt to ditch Intel (INTC +2.90%) processors for home-grown Apple-designed chips.
I don't think that's likely to happen in the near-future, but a case can seriously be made that Apple is moving at a faster clip than Intel in terms of ultra-mobile chip performance. Allow me to explain.
Meet Kaby Lake and A10 Fusion
This year, Apple's big new chip is the A10 Fusion found inside of the iPhone 7 and 7 Plus. Intel's new chip -- at least for ultra-thin notebooks and 2-in-1 devices – is a chip marketed as seventh generation Core (though I'll refer to it by its code name, Kaby Lake, as Intel's branding is fairly cumbersome).
The A10 Fusion is believed to be manufactured on Taiwan Semiconductor Manufacturing Company's 16-nanometer technology, the same tech that was used to build a portion of last year's A9 chips. Kaby Lake is manufactured on an enhanced version of Intel's 14-nanometer technology, known as 14-nanometer+, that was used to build the company's previous generation personal computer processor family known as Skylake.
So, in terms of manufacturing, neither Apple nor Intel gets a large generational leap (though Intel probably delivers a bigger jump here relative to what Apple had to work with).
However, what Apple has done in terms of the actual chip design is much more aggressive.
A10 Fusion sees huge design improvements, unlike Kaby Lake
With the A10 Fusion chip, Apple appears to have completely overhauled the design relative to the prior generation A9. The CPU cores look improved, the graphics processor looks new, and Apple was keen to tout the enhanced image signal processor (critical for the operation of the camera) inside of the new offering. I wouldn't be surprised to learn of additional goodies inside the A10 as details become public.
Kaby Lake, in contrast, is essentially Skylake implemented in an improved manufacturing process with some enhancements to the media engine to support more complex content.
To be clear: Skylake was and still is a really good processor for personal computers, and Kaby Lake should continue to be a leading chip family for personal computers. However, the generation-over-generation improvements that Intel delivered with this new offering simply weren't close to what Apple was able to deliver with the A10 Fusion.
Intel needs to run faster
Although Apple and Intel are not direct competitors -- in fact, Apple buys Intel's personal computer processors for its Mac lineup -- I do think that the advancements that Apple, and others in the smartphone chip space have made, should serve to push Intel to be more aggressive with the improvements that it delivers with each new processor generation.
The key areas that Intel needs to focus on are CPU performance, graphics performance, and integration of additional components (something that I covered extensively here).
According to respected processor analyst David Kanter with The Linley Group in a statement to EETimes, the reason that Kaby Lake didn't see a significant architectural improvement is that "by the time Intel discovered the yield issues with 14 [nanometer] and that they would continue for 10 [nanometer], they could not tell their architecture teams to tweak the CPU and GPU cores again."
The good news is that Intel's third generation of 10-nanometer product, known as Tiger Lake, has been in the works since at least late 2015, so Intel should have plenty of time to bring more substantial improvements to the chip.
