Everything, it seems, is a computer these days. Oh, sure, we call them “phones” and “cars” and refrigerators,” but each of these old terms is so stuffed with new things like software and semiconductors that they long ago stopped being “that thing I drive around.”
Don’t believe me? Take Ford’s recent announcement that it will slash production of its hugely popular F-150 truck by up to 20% this quarter due to a shortage of chips. Not engines. Not steering wheels. Not fuel systems. Chips.
Among the auto manufacturers, Ford isn’t alone–GM and others have also announced slowed production as they await chips. The question is whether these “non-tech” industries will feel sufficiently motivated to start building their own chips, as Apple and other big tech companies have. These mainstream enterprises have already started to get into software–is hardware next?
SEE: Hardware inventory policy (TechRepublic Premium)
Hardware is hard
We’re so far into the “software is eating the world” phase of corporate existence that we may not remember that it wasn’t too long ago when software wasn’t considered a strategic asset by most enterprises. Yes, every company wrote, rather than bought, most of its software, but it didn’t yield the competitive differentiation it does today. It’s become cliche, but today every company really is a software company, or trying to be one.
As enterprises build expertise in software, they’re increasingly tapping into open source communities to improve that software. Take Lyft, for example, with its service mesh, Envoy, now a CNCF project. Or Capital One, which has given rise to a number of open source projects, or has fostered the growth of others, like Cloud Custodian. When non-tech companies start to get involved in open source like this, it’s a sign that software has truly arrived as a competitive differentiator within the enterprise.
But not hardware. Not at the hardware infrastructure level. Why?
We’ve already seen movement in this area. Apple has increasingly shifted for its homegrown chips, first for smartphones and now for laptops. Microsoft, AWS, and Google have done the same, seeking better performance and lower costs. As detailed in a Wall Street Journal story (subscription required), these companies are building programmable chips, central processing units (CPUs), and more. Why? These “companies are searching for new ways to eke out better performance, not always measured in speed, but sometimes lower power consumption or heat generation.” These companies are finding they need custom chips, not the generic designs that chip companies like Intel have offered.
But these companies have something most enterprises don’t: Scale.
Yes, Ford sells millions of cars every year, for example, but it’s not yet clear whether it would make sense for it to create and produce its own chip designs cost-effectively (manufacturing can be done through chip fabs like TSMC). No, a company like Ford doesn’t want to be at the mercy of global shortages for generic chips (though it would remain at the mercy of manufacturing availability), but it’s also not necessarily well-positioned to build its own. The upfront cost for any of these companies is hefty, and as hard as it has been for them to hire great software talent, it’s even more difficult to find the equivalent in hardware talent.
As in software, open source may play a part. Open source hardware like Arduino and Raspberry Pi are teaching developers to “hack” hardware, while chip designs have found their way into the realm of open source as well. We’re nowhere near the level of maturity in open source hardware as we are in open source software, but that may simply be a question of timing.
After all, not too long ago the idea that software would be central to every company’s strategic value was farfetched–today it’s the norm, and open source is a key driver for that trend. Will hardware become the same? There are good reasons to say ‘yes,’ as well as ‘no.’ I’d love to hear what you think. I’m @mjasay on Twitter if you want to discuss it there, or below in the comments.
Disclosure: I work for AWS but the views expressed herein are mine.