The size of components in classical computers is shrinking exponentially, if trends continue soon the behavior of the components will be dominated more by quantum physics than classical physics. Researchers have begun investigating the implications of these quantum behaviors on computation. Surprisingly it seems that a computer whose components are able to function in a quantum manner are more powerful than any classical computer can be.
The study of quantum physics has shown that the behaviors of physical systems that understood from classical physics break down in systems of sufficiently small scale. The scale which these classical assumptions break down is the scale we are rapidly approaching with components in modern computers. At the current rate of miniaturization memory components will reach the size of about one atom per bit around the year 2020. At this size the storage for a bit cannot be expected to behave in a classical manner, and the size of an elementary particle is an absolute lower bound for the classical bit size. This bound does not hail the eventual end to advances in computing hardware. The same quantum effects which will prevent the continual miniaturization of classical computers may allow the development of a quantum computer, which relies upon these quantum effects.
At the heart of quantum computing lies parallelism. This inherent parallelism, combined with the exponential slowdown in any classical simulation of a quantum system make simulation of a quantum system an ideal candidate for speedup through parallelism.