I Didn’t Know It Was Loaded

"Companies tend to look for success in the bottom line."

-Anne C. Weisberg

When one searches the Thomas Register for companies that make integrated circuits, the number of companies that are in the Midwest is a fraction of a percent of the whole. What can be found is a small sprinkling of "fabless" companies that design semiconductors, but don't actually make them. Of course, very few small companies have the capital to invest in the multi-billion dollar manufacturing facilities that have football field sized production spaces so clean that that the only dust to be found is no larger than 300 nanometers in size.

It was not always so…

In July of 1968 a small company started with the purpose of developing a memory device based on silicon rather than the then prevalent magnetic memory.

Woven magnetic core memory bits

At the time, silicon memory cost almost 100 times more than the hand woven magnetic cores that constituted the memory elements of computers of this era. Facing various investment hurdles, and potential competition from other companies, the founders changed course and took the brazen path of producing the world's first microprocessor, the Intel 4004, which contained 2,300 transistors in the space of a postage stamp, cost an earth shattering $200 per chip, and executed instructions at the rate of 1 every 1/60,000^th of a second. (It would take more than 250,000 of these chips to match the performance of a typical processor in today's home PC.) Faced with financial success, the company went down the path of stuffing ever more functionality into their microprocessor family creating generation after generation of microprocessor.

Three years before the founding of Intel, Gordon Moore, then at Fairchild Semiconductor, who would later be one of Intel's founders, wrote in a paper for Electronics that suggested that it seemed as if the industry was doubling the density of components every year. He suggested that if this trend continued, the present density of 50 components per circuit might reach 65,000 components per circuit by 1975. Somewhere along the way, this observation and speculation became "Moore's Law" and the operational standard for the semiconductor industry. Today, it seems as if "Moore's Law" is the law as far as the electronics industry is concerned.

There are two other interesting things that happen as "Moore's Law" continues down its economic course. First, the cost of the factories that build these devices increases by a factor of 2 every year or so. Second, the size of the features used in building these devices is currently in the 130-nanometer range, and shrinking. Intel is investing a few billion dollars in the next generation factory, which will build features as small as 65 nanometers, which means features will be brought to market that are on the order of 300 atoms wide.

Assuming that this trend continues, there is a wall to be found. At some point in the not too distant future, the smallest feature size is about the same size as the smallest hunk of matter—the atom. When, and if, the feature size is the size of an atom, this is going to be a rather imposing physical problem, which limits how long "Moore's Law" can hold. Pessimists give it 5 years. Optimists forecast the end to "Moore's Law" not later than 2017. Anyway you look at it; there is a limit to how small you can make things out of atoms.

On the other hand, if you're a nanotechnologist and really don't care about computers that much, there is something rather interesting that Intel has been doing for the last 30 odd years. By constantly shrinking the size of components in their products, Intel has finally broken the hallowed 100-nanometer barrier and will be routinely manufacturing components in the clear realm of nanotechnology by the end of this year. This makes Intel the world's largest nanotechnology company.

What I like most about Intel is that they don't think of themselves as a nanotechnology company. Intel's chairman, Andrew Grove states "The two areas that our business focuses on, computing and communications, are the backbone of the digital infrastructure, and our products are the building blocks that makeup this infrastructure." Did anyone see the word "nano" in there? I didn't. Yet Intel has public plans to produce transistors as small as 20 nanometers by 2007.

The key point of this is that nanotechnology, for those who enjoy market success, is just another hurdle on their path to market success. Like overcoming unions, shipping strikes, and foreign competition, nanotechnology is just another one of those things that gets in the way that has to be managed and cudgeled until it bleeds profit.

In four weeks, we'll take a look at a different company that positively drips "nano" albeit without profits. Next week, we're going to enter the dreamland world of nano-possibilities.