Unfortunately, Intel has quite a bit of research to do before these new transistors can make their way into actual processors. There are a few shortcomings to today's technology that need to be overcome before we see high-speed, low-power integrated circuits on the scale Intel is anticipating.

In order to maintain a consistent charge throughout the silicon used to manufacture a processor, additives called "dopants" are included. As transistors shrink, the concentration of dopant atoms must also increase to sustain the charge. At a certain point, though, the concentration becomes too great, and the dopant atoms clump together. To combat this limitation, Intel is researching a combination of different materials to construct the transistors.

Another inevitable barrier that will need to be overcome concerns the quantum nature of electrons. As transistor gates shrink to the scale of two or three atomic layers, electrons begin to pop in and out of the gates, regardless of whether the gate is open or closed. While the solution for this issue is still being researched, the basic idea would be to utilize different materials for gate construction that would allow a greater number of atomic layers occupying the same amount of silicon – a material with smaller atoms, which would keep electrons from popping up on the wrong side of a closed gate (in theory, of course).

Thirdly, variations in the concentration of dopants within each chip become increasingly critical as transistor and etching processes scale down. It would be nearly impossible to distribute the exact same ratio of dopants and silicon throughout the millions of transistors on each chip wafer, and as manufacturing technologies become more complex, yields decrease and production becomes less economical.

On the topic of economics, Intel will also have to fight Mr. Gordon Moore's second law, which states that as integrated circuit complexity increases, so does the cost of manufacturing. For example, a fabrication plant built in 1965 cost close to $12,000. On the other hand, a facility set up today to produce wafers based on an .18-micron process costs nearly 2 billion dollars. No, we won't accept "inflation" as a reasonable excuse for such a sharp increase in cost.

So now we know that Moore's law potentially lives on. We also know there are quite a few barriers that need to be broken before manufacturing technology can advance. So what is the potential of a 10GHz processor with 400 million transistors?

One possible application is real-time speech computation similar to today's voice recognition software. However, instead of merely displaying the speech on a page, the computer would be able to obey commands or translate to another language in real-time. Intel would also like to see these new technologies utilized to automate everyday home and business tasks, like shopping or security system control.