One more about the end of Moore's Law.

The basic assumption underlying the Moore's Law curve is standardized engineering practices, namely the silicon-based planar integrated circuits, which use silicon crystal growth, photolithography, various deposition techniques, wire bonding, etc. The standardization of engineering practice since the late 1960s has led to a proliferation of small, specialized firms that not only make chips, but also make the equipment and supply the materials in the manufacturing process. For the former, think Intel, Motorola, Texas Instruments, and other foundries in Asia (TSMC, for example); for the latter, think Applied Materials, Teradyne, KLA-Tencor, Lam Research, Electroglas, Kulicke & Soffa, etc.

Scholars have argued that the transition from vertical to horizontal structure was due to rapid technological change, which made it difficult for vertically-integrated firms (IBM, RCA, GE, etc) to keep up. Lewis Branscomb, a long-time IBMer and professor of the Harvard Kennedy School, pointed out in a 2004 lecture that "the speed required for innovation can't be done within a vertically integrated institution." The industrial ecology that we see today in Silicon Valley was a result of that transition. And, the industrial ecology currently hinges on the standardized engineering practices that is widespread around the world.

Now, what will happen when the standard platform undergoes drastic change, as in the "end of Moore's Law"? Will the industry temporarily return to its vertically-integrated structure, before returning to the standard equilibrium? What if the entity that comes up with the breakthrough technology decides to keep it for as long as it can? What if it decides to make the manufacturing equipment in-house?

The "end of Moore's Law," should it will occur, would require an extensive organizational adjustment across a wide array of institutions that comprise the "extended semiconductor industry," broadly defined.