The commercial applications gap

These days too many seminars, papers, and grant applications begin with great claims about the potential commercial applications of the research being discussed.
We should be skeptical about any hype concerning technological applications of basic research in materials science.
There is a big gap between a commercial device/material and what you can do in the lab with millions of dollars worth of equipment on a milligram of a material or a single electronic device.

It does not matter whether it is a photovoltaic cell, a catalyst, or a superconducting wire. All of the following demanding criteria must be met. Furthermore, it must be better than any existing technology and any competitor on most of these counts.

Cheap to manufacture.
Scaleable to mass production.
Durable. Often on the scale of years or decades.
Reliable and reproducible. Devices, whether batteries or computer memories, must work all the time.
Healthy. Not expose the user or manufacturer to toxic materials.
Use materials available in abundance (silicon, water, ...) rather than scarce ones, such as some rare earth elements.
Environmentally friendly.

Thanks to Tanglaw Roman for emphasizing these issues to me.

This post was partly stimulated by re-reading the front page of The New York Times from March 20, 1987, which features an article Discoveries bring a `Woodstock' for physics. The article describes the famous session on cuprate superconductors at the 1987 March meeting of the American Physical Society. It is worth reading to see how so little of what was promised then has not happened (unfortunately).

Can you think of other criteria that new technologies must meet to be commercially viable?