Over the years, there has been a lot of talk and speculation that Apple (AAPL -1.22%) would transition its Mac line of personal computers away from Intel (INTC -2.40%) processors and toward internally designed processors.
The truth is that Apple has made substantial progress in building very low-power, high-performance processors for its iPad and iPhone lineup. Intel's highest-performance desktop and laptop processors still perform better than what Apple has been able to endow its latest iPads and iPhones with, but those chips consume substantially more power.
Image source: Apple.
In fact, in the one product category where Intel's parts are required to operate at iPad- and iPhone-like power levels -- the 12-inch MacBook -- Intel appears to have fallen behind.
The evidence
In the Geekbench 4 benchmark, which measures the CPU performance of a system, Intel's best 4.5-watt processor, the Core i7-7Y75, achieves a single-core score of 4540 and a multicore score of 8173. This is the best score that I could find in the Geekbench 4 database for this chip.
Although I haven't personally tested the A11 Bionic chip inside the iPhone 8/8+/X yet (my iPhone 8+ is scheduled to arrive soon), the best single-core score that I could find in the Geekbench 4 database was 4275; that same phone also yielded a score of 10467.
As far as the single-core score goes, the Intel chip wins some subtests, while the A11 Bionic chip wins others, making it a wash between the two devices in my view. The A11 Bionic wins in multicore, but it's worth keeping in mind that it has six processor cores -- two high-performance cores and four high-efficiency cores (to use Apple's terminology).
The Intel chip has just two high-performance cores that have a relatively wide dynamic range of operation.
So, in terms of the single-threaded performance that Apple can deliver in a mobile form factor, it has caught up with Intel and surpassed it in multicore performance.
The problem
If Intel can keep rough parity with Apple in these very low-power envelopes, then I don't see a compelling reason for Apple to switch away from Intel in these systems. Apple would have to do significant legwork in terms of both chip design as well as in software ecosystem migration to transition from Intel Architecture-based chips to its own ARM-compatible processors.
The Mac business just isn't that big a part of Apple's overall business to make such a shift worth it for a minor performance bump.
The problem, though, is that if the gap starts to get too wide between what Intel can deliver in a highly frugal power envelope and what Apple can deliver in a similar power envelope, then I'd imagine that Apple would be increasingly tempted to build its own Mac processors.
If Apple builds its own chips and they're roughly on par with Intel's, then that's not exactly a compelling selling point -- a lot of work to replicate what can be purchased off the shelf, and manpower diverted away from iPhone and iPad chips (where Apple's differentiation is clear and much more impactful to its top and bottom lines).
However, if Apple can build demonstrably superior chips for its Macs than what Intel can cook up, then Apple's top brass might think that moving over to its own chips for a clear marketing advantage would help its business.
In fact, don't be surprised, when Apple announces its next iPad Pro powered by what will presumably be called the A11X Bionic chip, that Apple achieves a decisive CPU performance lead over Intel in the world of ultra-low power processors.
Intel's way out
I think Intel can successfully retain its position in Apple's Macs, but it might need to fundamentally rethink its approach to building processors targeted at power levels appropriate for the 12-inch MacBook.
Right now, Intel's attempt to build fundamental processor technologies that scale from chips that need to go into thin, light, and fan-less notebook computers all the way to chips that can go into enthusiast gaming desktop processors may be yielding suboptimal results.
Either something must fundamentally change or Intel faces increased risk of losing Apple as a customer in at least some of the latter's product lines. Such a loss wouldn't be catastrophic to Intel's financial results, but it would be a significant impact that would also have the side effect of dramatically diminishing investor confidence in the company.
