More on ARM-Based Macs

Last month I wrote an article titled “Why Apple Will Switch To ARM-Based Macs”, and Stephen Hackett responded to it with a piece titled “Follow-Up On The Possibility Of ARM-Based Macs”. In his article and in a tweet written on the same day, Hackett pointed out that I own stock in both Apple and ARM.

I interpreted that as an insinuation that I wrote what I did because of my stock ownership. That isn’t the case, and I was offended at the suggestion.

But after thinking about it some more, I realized that Hackett was 100 percent right to question my motives. If he’d written an article arguing that Apple will switch to ARM-based Macs while owning shares in both Apple and ARM, I’d question his motives, too.

His article and his tweet made me realize that I need to be more transparent than I’ve been. As such, I’ve added when, why, and at what prices I bought my shares of Apple and ARM, and why I continue to hold those shares today, to the left-hand side of this site. If my financial holdings change in any way, I will update the sidebar to reflect those changes.

Now, on to Hackett’s article:

Even if Apple could ship a ARM-based Mac [sic] that’s as fast x86 Macs [sic], Apple would be sacrificing Windows compatibility, which is a huge deal for many business users, including yours truly.

He’s right that Windows compatibility is a big deal for many business users. But if Apple switched to ARM-based Macs, the company wouldn’t be sacrificing Windows compatibility — it would be sacrificing native Windows compatibility. There’s a big difference. ARM-based Macs would be able to run Windows, just not as fast as Intel-based Macs can.

Hackett also points to an argument John Siracusa made against ARM-based Macs in an episode of Accidental Tech Podcast. To wit: Because Intel’s fabs are more advanced than everyone else’s, Apple’s ARM-based chips wouldn’t perform on par with the Intel chips they’d be replacing.

(For those who don’t know, the chip creation process can be broken down into two separate and distinct steps, and each independently contributes to the overall performance of the resulting processor. The first step is chip design — figuring out what features the processor will have and how it will work. The second is manufacturing — turning a file that exists on a screen into a physical product you can hold in your hand. A 7 out of 10 chip design combined with a 10 out of 10 manufacturing process results in a processor with an overall score of 17 out of 20; a 10 out of 10 chip design combined with a 7 out of 10 manufacturing process also results in a processor with an overall score of 17 out of 20. Strength in one step of the process compensates for weakness in the other.)

What Siracusa missed, however, is that Apple has consistently produced chip designs that are so incredible they negate Intel’s manufacturing process advantage. The 28nm A7 chip in the iPhone 5S consumes less power than Intel’s 22nm tablet chip, yet the two processors perform similarly. The Intel chip is manufactured with a more advanced technique, but the A7’s superior design compensates for it.

The A8 is a similar story. Intel’s process lead is supposed to give the company an advantage in transistor density, but Apple’s A8, manufactured on a 20nm process, is denser than Intel’s 14nm chips. The A8’s superior design negates Intel’s advantage in process technology.

Mobile chips like the A8 are harder to design than desktop chips because they have more onerous power consumption and size constraints, and Apple’s mobile designs are so incredible they negate Intel’s fab advantage. If Apple can design mobile chips — which, again, are harder to design than desktop chips — far better than Intel can, it can definitely design desktop chips powerful enough for the Mac. Apple’s ARM architecture license allows it to, and the company has incredibly strong incentives — better user experience, more profit, more control — to actually do so.

Apple will switch to ARM-based Macs. It’s just a matter of time until the transition is announced.