A lot of smart people are saying that Apple is going to announce its own CPUs at their upcoming WWDC Developer Conference. Having spent many years working to build support for an ARM-based CPU, we want to touch on the significance of this and look at what else this change may bring.
For almost 40 years, Intel’s x86 has been the dominant force in the most important computing market in the world – CPUs for computers. Intel powers the bulk of the billion or so PCs in the world today and billions that preceded them. This has been an incredibly lucrative franchise making Intel the largest chip company in the world.
The advent of smartphones in 2007 cracked a big whole in that position, but for the past 13 years Intel has continued to reap huge profits. Intel completely missed out on mobile, despite spending ~$20 billion to get back in. They survive because PCs haven’t gone away.
However, the growth of mobile, and importantly growth of web-based software (aka the Cloud, aka Software as a Service, aka mobile apps) has radically altered what we do with computers. Successful software no longer has to be optimized for x86 Intel chips, it has to be optimized for running in the cloud and displaying nicely in a browser or in an app.
So if (when) Apple announces its own ARM-based CPUs it will mark a very important milestone away from Intel.
The interesting part of this is not that a large laptop maker is moving to ARM, it is rather that it has taken this long. There have been so many cracks in x86’s dominance, the fact that the whole edifice still stands speaks to something deeper taking place.
Intel’s dominance in CPUs is less about chip performance than it is about software ecosystems. People still use Intel-based PCs because they bring with them 40 years of software investment around them. This spans the gamut from desktop applications to device drivers to low-level firmware and BiOS. Replicating these is hard – there are hundreds of companies involved, and there is considerable legacy, backward compatibility required.
Apple can lead the way on this because so much of its ecosystem exists solely under its control. Their entire operating system is known for being a closed sandbox, something that was never possible for Intel which had to work across dozens of hardware companies.
Nonetheless, we think it is possible that other PC makers have to be looking at designing their own chips now. It is not clear whether a Dell, HP or Asus can design their own chips, but Samsung and Huawei can. The Windows PC makers for years have been trapped in No-Profit land squeezed between Intel and Microsoft. They now have the potential to prise themselves free from one half of that.
Apple has demonstrated in smartphones that designing your own processor can be a massive competitive advantage. To the point that all the other serious phone companies are now working on their own chips. We would argue that for a large segment of PCs, the same could be true. Admittedly, the global enterprise IT complex is unlikely to shift its knowledge workers away from Windows any time soon, but we can also think of large segments of that market which would probably benefit from moving to lower cost non-Intel, non-Windows machines. Take the example of call centers, or international border crossings. Every work station in those places has a PC today. All of which is overkill for the way the device is used. Similarly, in China there is a big national effort to move away from Intel and Microsoft. There are likely many, many other markets where a stripped-down compute device is more than sufficient.
This has been technically possible for years, but the sticking point was all that ancillary middleware. Apple is likely about to demonstrate that much of that is no longer necessary. They are not going to be the last to test this theory.