With all the gloom and doom facing the semiconductor industry and especially with the "end of Moore's Law" coming up soon as many experts predict, let's look at several facts relating to the unbelievable ride the industry has had for the last 40 years.
Moore believed that scientific advances affecting semiconductors could be crucial to economic growth, because an extensive range of applications would be found for more powerful devices in industry, government and national defense. He thought that it would depend on a tradeoff between the pace of scientific advance and the costs of producing more powerful devices.
As the IC industry is progressing and many layers on the silicon chips are approaching atomic scale, we might be facing a insurmountable wall that could hamper the future development of chip technology. Many changes took place in the industry as copper replaced aluminum and CMP was introduced. High k hafnium oxide replaced the traditional silicon dioxide as the basic building block of the transistor (gate), mobility enhancement was obtained by stressors (either dielectric film or selective epitaxial growth in the source/drain area) and of course the tri-gate process was recently debuted by Intel. The transistor of advanced processes today (32-nm and below) looks dramatically different from the 130-nm transistor of 10 years ago.
The end result of all these is: yes we have better, faster and more advanced devices ( as well as manufacturing facilities) but the cost of manufacturing has gone up dramatically as shown in Figure 1.
If Moore's Law comes to an end this decade, we might see engineers think of smarter ways to make devices smaller and smarter, perhaps maximising the computational power of existing chips. The electronics might hit a snag, but I think it gives lesser advanced societies time to catch up while we think of ways to innovate again.
Simon - http://www.starrausten.com
Using depreciated Fab is so appealing when the price of new fab is becoming redicuously high so this trend will continue. Now regarding analog stand alone vendors I think there is room for growth in the future especially in the smart grid, solar convertors, electric and hybrid cars etc.
Analog chips for RF and PMU paired with ditigal ICs on handheld devices is actually diminishing with SOC initiatives even though the market for these devices is skyrocketing, in fact because they are skyrocketing. As such I would argue the analog pure-play vendors are in danger of losing their shirt. As for Moore's Law, 3D and using already depreciated older fabs, all valid points until you need 32 cores.
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.