As the scale of computations become large and as both people and machines demand communication, collaboration and commerce at the speed of light, rapid fluctuations in the demand for computing performance and fluctuations in available resource pools, both make it necessary to respond fast and readjust the computation structures and associated resources so as to not disrupt the user experience or the service transaction. Current Information Technologies from their memory-starved, server-centric, low-bandwidth origins from von Neumann's stored program control implementation of the Turing Machine are evolving with new architectures to meet the demand for scale and speed. In this paper we discuss the evolution of current IT to cognitive IT, where computing processes become self-aware of their resource requirements in real-time and seek to adjust them from a global knowledge of available resource pools and their provisioning processes. This is transforming the current state of the IT as we know it to a cognitive IT.