For decades semiconductor developments have been driven by Moore’s Law productivity gains. This led to extremely fast digital processors, increases in bandwidth and huge memories that boost the productivity of PCs, mobile phones and other applications demanding heavy data traffic and storage. However, the economics of the IC industry as well as developments in society, are creating a paradigm-shift in the semiconductor world.
Our industry is confronted with sky-rocketing cost levels for system-on-chip development in advanced CMOS technology, whereas, ongoing shrinking of physical dimensions will eventually cause Moore’s Law to come to an end.
At the same time, the added value perceived by consumer demand points away from ever more productive consumer electronics and into new smart products. A globally ageing population demands innovation in medical devices and food safety. Raising environmental awareness demands smart “green” solutions. And in transportation IC innovations help to bring safety, to overcome traffic congestion, and to deliver real-time information, entertainment and services to people on the move.
Collaborative R&D models to change
These and many more applications are typically realized by integration of existing CMOS technologies with so-called “More-than-Moore” technologies, such as analog/mixed-signal, high-voltage, and ultra-low-power.
This shift towards multiple-technology devices will also affect development in design and architecture methodologies, modeling and characterization, and system architecture. Ultimately, current models for collaborative R&D will change, and new ecosystems will arise.
The paradigm shift from generic CMOS-driven technology convergence towards technology diversification provides semiconductor manufacturers with new means for differentiation and – much welcomed – new business initiatives.
Companies with experience in multi-technology product development, and known ecosystem drivers, such as NXP, are well-positioned to take their share of new emerging markets.
As a relatively young industry, the drive for semiconductor device manufacturers has historically been on productivity gain, since Moore’s Law first predicted the number of transistors on a given area of silicon would double every 18 months. This law has held true ever since and the productivity ratio has also driven transistor costs down, to a point where an affordable system-on-chip for consumer applications today may contain hundreds of millions of transistors. This tremendous productivity gain has also led to the wealth of electronic applications that we see around us.
Are we now witnessing the end of Moore’s Law? The answer to that must be no, but just as transistor density in production has increased, so too has the cost of developing the process that enables it. The result is that the most cost-effective IC solution today is no longer by default based on the smallest – and most expensive – CMOS technology available.
A parallel in aviation mirrors this shift from performance- to economics-driven development. After decades of speed increases for planes the revolutionary Concorde failed to become the new standard for passenger transportation through the air, despite its record-breaking top speed of around 2000 kilometers per hour.
The ability to travel from Paris to New York in only three hours was expected to turn the Concorde into the new standard. But it appeared that economic rules prevailed over sheer performance: extreme fuel consumption caused fares to rise to unacceptable levels, and Concorde has ended up in the museum.
The average speed in aviation found its economic equilibrium at around 1000 kilometers per hour and has remained unchanged for many years now.
But did it mean the end of innovation in aviation? On the contrary, the scope of innovation was redirected to add other value to passengers, crew, and airlines than just speed. Planes became safer, more comfortable, more cost effective, more fuel efficient, and less noisy.
NXP foresees a similar process to take place in the semiconductor industry, where the scope of innovation will shift from productivity to quality-of-life gains. As in aviation, there’s more to life than speed.
So, Moore’s Law, which dominated process technology in semiconductors for decades, will remain the driver in a limited number of high volume areas. But it may cease to exist the moment the boundaries of physical ability or economical feasibility are achieved.
Trends in society all create significant market opportunities for companies who provide state-of-the-art ICs in areas such as energy-saving in consumer electronics and lighting, medical diagnostics equipment, food safety, and traffic management systems.
Growth of these markets is not hindered by the economic downturn, since most of the chip innovations offer great return-on-investment; the product costs are often earned back easily on energy saving, less costs for medical care, less economic damage through traffic congestion. This way, real need-driven innovations in semiconductors can often pay for themselves.