Apple takes a bite out of Intel
There has been much hoopla around Apple’s announcement that they will replace the Intel chips in Macs with their own, but it may not be clear why to many consumers. In fact, Apple’s outstanding chip design is an integral part of the iPhone’s success. As Apple highlighted in the WWDC 2020 keynote, the A series chips used in all iPhones have provided reliable speed and battery life improvements year over year. Bringing in-house CPUs to the Mac lineup will likely bring greater performance and definitely leverage substantial synergies.
First, some background on Apple’s past experience with designing processors. In recent years, new iPhone announcements have been joined by flashy headlines of outlandish performance numbers. An iPhone outperforming a MacBook? These claims would be quickly ignored as a quirk of the testing, or limited to short bursts of computing. Still every year, A series chips have gotten significantly better, while Intel has to jump through hoops to feign progress.
Apple’s iPad performance is even more impressive. They have consistently beat out the entry level MacBooks and even make some MacBook Pros look underpowered . From tablets it’s not hard to imagine creating laptop processors. The rumors have been floating around for years, but all the pieces have finally come together in 2020.
Every report mentions that Apple Silicon will bring better performance. Better performance is kind of vague, so I want to answer why Apple would want to make their own chips for Mac products. Intel’s performance gains have been stagnant for a couple of years, but AMD’s improvements have been repeatedly blowing them out of the water. Couldn’t Apple switch to AMD for better chips? Plus, companies like Intel and AMD who have built laptop and desktop chips for decades seem to prove that desktop chip design is really hard. This leads to the first point of Apple Silicon: switching to an ARM architecture.
You may have heard of ARM in relation to smartphones and tablets. That does not mean they are limited to mobile applications. The architecture related to most laptops and desktops is x86, created by Intel in 1978. x86 is what Intel and AMD use for their processors. Also, it was basically the only ISA for decades until smartphones became relevant, and ARM began to take over. Despite the popularity of x86, there are concerns with the architecture. x86 has a very diverse development history. Intel created the initial architecture, but multiple other firms have expanded on x86 and created their own extensions. This contrasts with ARM, which does work with partners, but ultimately they control ARM updates.
The contrasting development histories mean that x86 and ARM also contrast in terms of their complexity. One way to measure complexity is with the number of instructions that each supports. There are a lot of ways to count instructions, so I’ll be general. x86 supports thousands of instructions, while ARM has around 50. This staggering difference means that x86 processors generally have to be more complex just to function properly. From that simpler base, ARM processors can be optimized for better efficiency, while maintaining performance.
In addition to the potential hardware gains, switching to ARM means that suddenly every iOS and iPadOS app is available for the Mac. Apple already announced this at WWDC, while giving developers to option to opt-out. This will partially depend on developer cooperation, but the implications are still compelling. Imagine all the apps you use on your phone, appearing on your laptop. These apps will generally be optimized for relatively limited system resources (like you find on a phone), so they will be efficient and super fast on a laptop. For example, instead of always having a tab open for Asana, you could just open and close the app when you want to check something. Now extend that to how many ever services you keep open on your laptop currently, and the savings add up.
Combine that intrinsic simplicity in ARM’s instruction set with Apple’s experience building ARM processors and wealth of ARM-based apps, and the switch makes a lot of sense. This congruence only increases when you consider the further synergies with Apple’s in-house production.
The biggest advantage is simply the coordination possible by having CPU design in house. Previously, Apple had to plan their Mac releases around Intel’s schedule. You can’t refresh a laptop if the refreshed processor isn’t available. The benefits of controlling the processor timeline are obvious from the iPhone’s rigorous and consistent upgrade cycle. Everyone knows a new iPhone will be out each fall. The greatest deviation from the schedule is an occasional lower priced iPhone in the spring. It seems ordinary now, but it’s actually incredible that Apple is able to make new chips every year with significantly improved performance, not to mention the rest of the phone. Intel certainly hasn’t delivered consistent or much improved chips. That isn’t the only reason Mac updates have been so sporadic, but it definitely is one. I think in the near future, Mac hardware updates will be similar to iPhone ones. Yearly Mac refreshes may be incoming.
Another great benefit comes from Apple’s vertical integration. Currently, macOS is tailored to Intel processors. If there is something Apple doesn’t like about the chip design, they are out of luck. They are frankly a small partner relative to the PC market and much of the enterprise market. With Apple Silicon, they can tailor the chip design to their own goals. Then the operating system and chip design can meet in the middle, instead of the OS being forced to accommodate Intel’s design decisions. Similarly, Apple can focus the scope of their chips, because they have a narrow use case. The product lineup of MacBooks and desktop Macs are limited, so the chip design can be aligned to meet specific demands and nothing more. Intel on the other hand has to make decisions to satisfy tens of vendors from Apple and Microsoft on the high end, to budget players like Acer and Asus. With the ability to make a narrow range of high-end chips every year, Apple may be able to compete fiercely with Intel’s best very soon. Knowing Apple, they might have waited until that level was achieved before making the announcement.
Apple Silicon has garnered a lot of attention for a good reason; it is going to launch a whole new era of Macs. The move to ARM takes advantage of their experience creating ARM processors, and the inherent efficiency gains. It also leverages the plethora of ARM apps already published on the iOS and iPadOS App Stores. Creating their own processors allows for better planning, leverages Apple’s strength in vertically integrated computers, and provides simple opportunities for growth including increased profit margins. Taking on the challenge of even more processor design is a big announcement, but Apple has shown they are more than competent in that industry with their phones and tablets. With the rest of these benefits on top, Apple can dramatically increase the value proposition of a Mac device.
