Where the next big push will come from in semiconductors and why.
Progress in electronics has always been about combining more functions into devices and making access to information more convenient.
This is what drove the PC revolution in the 1980s, when centralized data was made available on desktops, and it’s what drove the notebook PC revolution in the 1990s as computers became untethered from the desktop, as long as you could find an Ethernet connection (or a 14.4 modem). The adoption of the Internet by corporations, starting in the mid-1990s, completed that picture by breaking down the last big barriers for ubiquitous communication.
During the past couple decades, the push has been on shrinking the device and improving connectivity everywhere. If you haven’t noticed it for a while, think back on how many dropped calls you used to have. And then consider what can be done with a mobile device now versus even five years ago. Add to that a better supply chain there is more that can be done with a mobile device, and the kinks have been wrung out of the supply chain to the point where smartphones are replacing basic cell phones even in developing economies. While this is good news from a market penetration and device replacement perspective, it’s also a clear sign that the mobile market is maturing the way the PC market has matured.
Throughout the 1980s and 1990s, the majority of chips being sold were PC processors. Over the next couple decades, it was applications processors for mobile phones that drove the market. We’re now at the point where something else has to take its place, which begs the question, what’s next and why?
The answer will likely come from a variety of markets rather than one, each of which may be enormous—self-driving cars, medical, industrial, and wearables. And while they may all loosely fit under the umbrella of IoT or IoE, the reality is that these markets are likely to develop independently and uniquely with a common connective thread for communication. What’s important to note, though, is that each does share some of the same traits.
For one thing, they are all based upon combining more functionality into one device. Wearables will likely become a huge market once battery life is improved and functions such as body chemistry sensors—heart attack, stroke and insulin highs and lows—are fully functional. Cars will evolve to the point where they are much more capable of accident avoidance, even if they aren’t immediately fully autonomous, through the incorporation of more sensors and logic. Medical instruments, ingestible communication devices, labs on a chip are all combining functions that will allow monitoring to be real-time, perpetual, with much more meaningful results. And industry is in the process of adding sensors into everything from valves to stockrooms to improve efficiency and reduce down time.
All of this bodes well for semiconductor volume and much more intelligent design. Whether there are more complex SoCs everywhere, or whether there are simply more sensors feeding into more centralized edge-of-the-network processors and cloud operations, is an ongoing debate. But the bottom line is that SoCs will be needed in more places as the definition of systems expands, and they will be much more complex as more functionality is added into these bigger systems.
This is one of the reasons there has been so much consolidation across the semiconductor industry lately—NXP buying Freescale for $16.7 billion, Avago buying Broadcom for $37 billion, GlobalFoundries buying IBM Microelectronics, Intel buying Altera, Tsinghua bidding $23 billion for Micron, along with a series of security company acquisitions by ARM, Synopsys,and others. Big companies are placing big bets on the next big things, and they’re buying up the pieces they will need to tackle those markets—customers, expertise and infrastructure. So take a deep breath, because if they’re even partially right, the next mad dash is about to begin across multiple markets.