The future of electronics is increasingly being shaped by two major trends: mobile computing and the “internet of things.” The pervasiveness of mobile is fairly obvious, with 1.2 billion units expected to ship in 2014. The internet of things is less obvious, but slowly becoming a reality. The idea is that all objects in our environment are equipped with sensors and identifying devices and connected to the internet. And I’m talking about everything, from buildings to freeways to food containers to medicine.
When the internet of things becomes a reality, companies would not run out of stock, as involved parties would know which products are required and consumed. Mislaid and stolen items would be easily tracked and located, as would the people who use them.
At the recent Common Platform Technology Forum — produced by Global Foundries, Samsung and IBM — Simon Segars, executive vice president and general manager of the physical IP division at ARM, spoke about the impact of the internet of things and mobile computing on the way electronics are designed and used. “Microcontrollers and sensors are getting embedded into pretty much everything we interact with,” he said. “You’re going to need very small and very power-efficient technology, and a power-efficient wireless network to bring it all together.”
Segars said changes in computing requirements will transform how chips in servers, wireless communication networks and mobile phones are designed. “It’s the case in pretty much any form of electronics, if you know what you’re doing, dedicated hardware is the best way of saving power. This is why you have a video engine and a graphics engine in your phone — because it’s very expensive to do it on a general purpose computer. The same holds true for servers. If you know what you’re doing, you might as well build yourself some dedicated hardware to do it,” he said. Servers of the future could potentially be powered by SoCs instead of dedicated, big processors. “This is a great thing for the fabless industry because the fabless world knows how to make SoCs really well,” he said.
Similarly, rather than get “fixated” by the apps processor in mobile phones, Segars said it’s important to pay attention to the other chips used, such as smaller control chips used to manage battery power or the touch screen, for example. “These aren’t necessarily manufactured on the most leading edge digital process,” he said. “These are using older, more mature processes which can drive higher voltages. There’s a need for continual evolution on that kind of process technology, because it’s going to be required for a long time.”
As far as the “internet of things” goes, Segars said it’s a “bit like science fiction right now,” but it’s probably going to happen. Sensors will be built into buildings, the freeway, and into the environment in general, detecting what is going on in the world. “It’s going to generate just a phenomenal amount of data which is going to back to all those servers, and it’s going to need processing and be delivered to you in a way that’s actually useful so you can make some decisions. All of this is driving the demand for semiconductors, driving the demand for lower power electronics and driving the demand for future higher bandwidth wireless devices between all these devices,” he said.
Phones are now outselling PCs and are the primary mechanism for people engaging with the outside world. Segars said this is driven by the technology innovations. “By continually driving power consumption down, we’ve been able to come up with these really small form factors, devices that are always on and always connected to the outside world. It’s really been quite staggering,” he said.
He noted that, while the technology has evolved, the industry behind that technology has evolved a lot as well. In the 70s, companies were vertically integrated. Today, the industry has desegregated from companies that did everything to these tiers of companies who have specialized. “The benefits of that specialization have been improved economies of scale and companies that have been profitable in their own right without having to do everything,” he said.
He also commented on the need to look at the supply chain in a new light. “It’s easy to think of this as a stack, from somebody doing design down to somebody doing manufacturing and building equipment, but it’s much more complex than that. In order to not lose efficiency as you fragment into this desegregated world, it become very important for companies to collaborate in what they’re doing. If you own everything, the good news is you can tweak absolutely everything because it’s all yours. But that’s very expensive. If you don’t own everything, you have to rely on all of these companies in your supply chain to actually work together and drive out potential inefficiencies.”
According to Segars, a better way to think about the industry today is a circle where all these companies in different industries are working together to provide an overall complete solution. “The really smart people have realized they have to share a lot of information to make what they do as best as it can possibly be to feed into this supply chain so the end product can be the best it possibly can be. “
Segars gave a tip of the hat to the high costs of building new fabs and to conduct R&D. Earlier in the forum, Samsung’s Ana Hunter put a price tag of a new fab at $7.5 billion. Segars help up EUV as an example. “Before EUV really gets unleashed on the world on a production scale, the R&D costs for EUV will be well over $1 billion.”
The real question though is whether the semiconductor can continue going down the same path given these high costs and approaching physical limits. “We’ve seen over the history of the semiconductor industry the costs of transistors go down, roughly halving about every 18 months for decades. These costs are going non-linear. The concern is that if we are unable to do anything about that, if we’re not able to continue to drive out inefficiencies, then you have to ask yourself the question, ‘Are transistors going to get more expensive?’ If that’s the case, then that really disrupts how the semiconductor industry has evolved for many, many decades,” Segars said.
Simon Segars joined the board in January 2005 and was appointed EVP and General Manager of the Physical IP Division in September 2007. He has previously been EVP, Engineering, EVP, Worldwide Sales and latterly EVP, Business Development. He joined ARM in early 1991 and has worked on many of the ARM CPU products since then. He led the development of the ARM7 and ARM9 Thumb® families. He holds a number of patents in the field of embedded CPU architectures.