What used to be a safe bet isn’t so safe anymore, but that isn’t all bad.
Ever since the introduction of the iPhone in June 2007, increasingly advanced SoCs have dominated the semiconductor supply chain, from tools to design houses to foundries. Android’s introduction in 2010 only cemented the market. Together they created massive demand for power-efficient chips that were dark most of the time, feature-rich, and which could respond within milliseconds to any command. And more important, they allowed people for the first time to go on the road and not bring along multiple devices.
This was like unplugging an entire civilization. Demand exploded, almost singlehandedly pulling the semiconductor industry (and most of the world) out of a deep recession. That, in turn, allowed many other sectors to gain a much firmer footing, ranging from automotive to consumer electronics. Military has picked up, as well, as cyberspace plays an increasing role in warfare. So has medical, although the pace of introduction for medical electronics is always far slower than anyone expects.
But as the market for smart phones begins to taper off, following the same maturing cycle that began hitting PCs early in the millennium—replacement cycles driven by aging electronics rather than killer feature introductions or massive capability improvements—questions are beginning to surface about how the semiconductor industry will need to change.
For starters, there is no immediate single threat on the horizon. Geopolitical and macroeconomic concerns are persistent and always have been. Individual markets fluctuate naturally, but none is gyrating too wildly. Even consolidation will abate eventually, probably leaving a sour taste in everyone’s mouth because companies inevitably will gorge on cheap capital and buy more than they need or should. But in general, there is no single worry that anyone can point to right now other than a question about what is the next big market driver.
While this may provide some jitters for certain segments, notably investors who have been used to watching steady gains at big systems companies, the biggest change is that a massive concentration of semiconductor content into the hands of fewer and fewer companies may be over. The number of chips shipped each year will grow, but with the maturing of markets such as smartphones (and its larger sibling, tablet computers) those numbers may not grow as fast in the same markets.
There are other bright spots, as well. The Internet of Everything is very real, although harder to measure because it is so diffuse. But the number of chips used in IoT gateways, edge node devices and servers, network gear and industrial applications is growing quickly. In the future, everything will be an IoE device, and everything will be connected. But it’s likely that all of these systems will not be powered by a handful of SoCs made by one company. They will be spread out among more companies, using more configurations of chips, running far more task-specific software.
This isn’t necessarily bad news for the semiconductor industry, but there will be some upheaval as markets reposition. Still, it does have some interesting implications for the future because the economics of billion-unit chip development no longer apply in this world.
In the next phase of the industry, lifespans for products will be much shorter. That means design cycles will need to be shorter and replacement cycles will be quicker. This is a net gain in terms of overall chips, but they will need to be built differently, using more IP, with much smaller foundry runs. Ecosystems will need to be looser and much more fluid to be able to adapt to changes, and R&D will have to shift from families of chips to cross-market platforms and faster ways to assemble different pieces.
Put in perspective, the shift from mainframes to PCs to handheld computers is nearly complete. In the future, everything will be a computer and a mobile communications device, which will have a pronounced impact on every part of the supply chain in every industry. Along with that will be lots of cross-market and cross-industry surprises, some good, some not so good, and enough to keep market analysts and economists busy trying to figure out the changes for many years to come.