After Intel: The Next Computing Revolution?

Apple's culture of secrecy rarely stops the rumours, and the topic of ARM-based Macs has attracted plenty of discussion within technology circles over the years. For many geeks, it was an interesting thought experiment. Last week, that thought experiment become a reality as Apple announced the transition of their entire Mac product line from Intel x86 processors to ARM-based Apple Silicon. This change will take two years, after which Macs will use upgraded versions of the A series processors currently used by iPhones and iPads. It was a big moment with enormous consequences for the entire industry, even if everyone knew it was coming.

When the first Apple developed processors debuted in 2010's iPhone 4, there was speculation that the tech giant's long-term aim was to see Apple processors inside their entire product line from the smallest iPod to the most powerful Mac Pro. The question was if such a thing would ever be possible, such was the capability gap between the Intel processors that power PCs and the ARM processors that power phones. More recent iPhones have answered that question. Technology has advanced a long way over the last decade, and the capability gap has closed dramatically, turning the latest iPhone into a computing powerhouse capable of out-performing some low-end PCs.

In 2020, no one doubts that Apple can develop a CPU capable of powering a Macbook or a Macbook Air. Indeed, Apple processors would probably make such devices faster by removing many of the thermal constraints placed on these devices in order to stop their existing Intel processors overheating. The big question is whether the same processors will be able to scale to the high-end Macbook Pros, let alone the Mac Pro workstations widely used for movie and music production in the creative industries. ARM-based processors developed by other companies are now being used to power servers and supercomputers, so in theory, Apple processors should be able to scale across their entire product line. Although, that doesn't necessarily mean they'll adapt to desktop use cases particularly when it comes to graphics performance in workstations. We'll find out more when the next Macbook Pro is released in October.

Customer Churn?

There is little doubt that Apple will lose some loyal Mac customers because of the switch to in-house processors. Every major Mac announcement over the past decade has been accompanied by cries of betrayal from the high-end professional users that once formed the core of Apple's customer base. This one is no different because there are software implications as well as hardware implications. Software developed for Intel Macs will need to be recompiled for ARM Macs, as the two types of processors aren't compatible with each other on a hardware level because they use different instruction sets. A translation layer called Rosetta 2 has been built into macOS to abstract this problem away from users, allowing software developed for Intel to run on new Macs. This should work well for most people because Apple have plenty of previous experience in this area from previous hardware transitions.

There is one group of Mac fans who will be left behind though: Windows users. The Rosetta 2 transition layer doesn't work for other OSes or software run in virtualisation. As such, the new Macs won't support Bootcamp and Parallels won't be able to run Windows or Windows software. That is a declining share of Apple's userbase but will affect some business users. The same compatibility challenge impacts open source developers too; a group who almost exclusively use Macs because of the similarities between Linux and the Unix internals of macOS. x86 versions of Linux won't run on the new Macs, although Parallels will still support ARM versions of Linux. There may be a hit to Apple's mindshare as the widespread use of Macs in tech firms is due, in large part, to their advantages when it comes to coding open source software. Some developers will be left with no choice but to switch to Linux or Windows, simply to continue working.

Windows on ARM?

Future updates to Windows may resolve this incompatibility issue, as an ARM version of Windows has existed since 2012. Microsoft's Surface Pro X already runs it. However, Windows for ARM is markedly inferior to the regular version. It doesn't run 64-bit software, a feature gap that Microsoft recently promised to fix, and has compatibility issues with some 32-bit software. These software issues may never be entirely fixed, because of the sheer amount of legacy code and backwards compatibility support built-into Windows that can't be ported over to the ARM version. But at some point, Windows for ARM will become good enough to run the majority of everyday software without issue. No only then it will be ready for installation on Macs, but also on a new generation of Windows PCs.

That day might come sooner than we think. At the moment, inadequate hardware is a far bigger barrier to the adoption of ARM-based PCs than the limitations of Windows. Apple processors have a substantial performance advantage over the ARM processors made by Qualcomm, Samsung and others for Android devices. That means Apple can release viable ARM-based Macs now, while PC manufacturers will need to wait until 12 months until their suppliers are capable of shipping ARM processors that can match an Intel or AMD CPU. Given the current state of Windows that's not a problem. ARM-based laptops should have a battery life advantage over ones with Intel processors, but high-end performance may still be limited in the medium term. In the meantime, it gives Microsoft breathing space to get their OS ready.

The End for Intel?

It also gives an already struggling Intel space to find a response. It's been a bad few years for the dominant force in desktop computing. A series of security flaws has seriously harmed the reputation of a company already struggling to keep up with their competitors. Their decades long dominance was built on massive advantages in manufacturing processes, that allowed them to compensate for serious thermal inefficiencies in their designs by building much smaller processors than their rivals. That process advantage has now disappeared, meaning that their rivals are now building much smaller processors as well as much more efficient ones. They've instead focused on increasing performance to compensate, helping them in the data centre market but allowing AMD to catch them in the PC market. Intel-powered PCs no longer have an advantage over AMD powered processors except at the very high end, while ongoing supply constraints have limited seriously harmed their relationship with PC manufacturers by limiting how many devices that Lenovo, Dell & HP can sell.

It's no wonder then that Apple have taken the opportunity to jump ship and sell ARM-powered computers. As always in the technology industry, they're not the first to do it, but they will be the first to make it work. In time, expect the rest of the industry to follow at least for low-end devices and thin laptops. There are still plenty of segments of the computing market where ARM-based PCs won't be viable for many years, so Intel still have a big market available to them but then so did IBM for a long time. IBM once dominated the high-end computer and data centre markets, until competing technologies gradually eroded their marketshare to the point that they now only sell servers to a declining number of large organisations that handle exceptionally high numbers of transactions per second. They're essential to the world's financial markets, but their hardware is seen as a legacy hardware technology with no relevance to the wider industry. There is a serious risk that Intel could follow that same path. Struggling tech giants are frequently labelled as the next IBM. The real question is who are the next Intel?