Prominent analyst and Apple-watcher Ming-Chi Kuo has told investors that he believes the Cupertino-based company will launch a system based on its own silicon rather than any chip from Intel in 1H 2021. Such an SoC would use an ARM architectural license and presumably be closely derived from the company’s various mobile efforts.
We periodically hear rumors that Apple is planning just this kind of departure. The first time I hit the topic was in 2014, but we had multiple leaks in 2018 and 2019 as well. As the years have passed and the performance of Apple’s SoCs have improved, it’s gotten easier to imagine the company launching new hardware built around ARM products rather than Intel. Intel’s 10nm woes and delays are a further example of why many people think Apple wants to go this route. The company would simplify its ecosystem, capture a larger share of the MacBook’s profit with its own manufacturing, and be able to enjoy the prestige of “crossing over” from mobile to desktop with a chip that could theoretically compare with the best Intel (or AMD) have to offer.
Kuo has been right on a lot of Apple predictions before — enough to make it worth taking him seriously on this one — but he’s also been wrong on some of the timing. In 2018, Kuo predicted that Apple would be launching such a system in 2020. Now it’s been pushed back to 2021. But in this kind of situation, being wrong on the timing isn’t the same thing as being wrong about the effort. There is any number of reasons Apple might delay formally launching a product built on silicon that’s very much in development, including wanting to wait until it has all the features it intends to introduce up and running at the performance it wants them to offer. The company could also have chosen to wait and deploy hardware until specific software features or capabilities are locked in for a given timeframe.
In 2014, the idea of Apple switching away from Intel was ludicrous given the state of the company’s CPU technology. In 2020, it looks a lot more plausible. While it’s difficult to compare performance between ARM SoCs running iOS and a Windows PC, the performance comparisons that have been made show Apple outpacing every other ARM SoC vendor in per-core performance and having closed the gap with Intel and AMD desktop CPUs.
The Emulation Problem (or Maybe “Problem”)
The big problem (maybe) for Apple is that software emulation isn’t a terribly effective way to offer effective performance in software written for a different architecture. Apple’s greater level of control over its own ecosystem would help it here, but the company can’t force third-party software developers to release new software versions for its ARM hardware or to release them on its own schedule. In the past, when Apple has made this kind of switch, it always switched to a much higher performing CPU design. Currently, the A13 Bionic is roughly on-par with desktop chips from AMD and Intel, but it’s not dramatically faster. It’s also not clear if Apple would be allowed to offer a 64-bit x86 compatibility layer — Microsoft’s own Windows 10 emulator lacks this feature.
The biggest question for Apple? How much it cares. The biggest difference between Apple in 2005 versus Apple in 2020 is that the Macintosh is a much smaller percentage of the company’s earnings. Apple may look at the x86 compatibility question and not care if the end result is less market share for Macs if it can capture a larger share of the profits per system. It might also be willing to take a hit on sales immediately after the switch (because reintroducing compatibility issues likely would hit sales), with the bet that it would win those customers back long-term with superior CPU engineering. Or, it may believe it can paper over the issue with the right emulation solution plus a good-enough ARM CPU core.
If Apple’s goal is to introduce a CPU core that competes across mobile, desktop, and workstation, it matters whether the company can build a chip to match 28 cores worth of Xeon. If it decides to pull back from those efforts or to only target part of that market, it doesn’t.
Here’s the full note, courtesy of MacRumors:
We expect that Apple’s new products in 12-18 months will adopt processors made by 5nm process, including the new 2H20 5G iPhone, new 2H20 iPad equipped with mini LED, and new 1H21 Mac equipped with the own-design processor. We think that iPhone 5G support, iPad’s adoption of innovative mid-size panel technology, and Mac’s first adoption of the own-design processor are all Apple’s critical product and technology strategies. Given that the processor is the core component of new products, we believe that Apple had increased 5nm-related investments after the epidemic outbreak. Further, Apple occupying more resources of related suppliers will hinder competitors’ developments.
The competitors in the last sentence are much more likely to be other smartphone SoC developers like Qualcomm rather than AMD. Qualcomm and Apple now compete to be early adopters, while AMD has been following on semiconductor tech. Apple, for example, will introduce 5nm hardware this year, while AMD will launch 7nm refreshes of existing products. By the time AMD is pulling the lever on 5nm, TSMC’s yields will be higher and Cupertino will itself be eyeing the next node.
- Apple Could Switch to ARM, But Replacing Xeon Is No Simple Endeavor
- Analyst: Apple Will Switch Away From Intel in 2020
- Why Apple won’t dump Intel x86 for its own ARM chips in MacBooks and the Mac Pro (2014)