AI and Challenges in recent times. Artificial Intelligence is built bit by bit and incorporated in to a machine to bring them to live . The intelligence required are to make the machines learn by them selves and progress with time. Expected to memorize, compute, decide upon any decisions required in a flash. The basic requirement for this to happen is High Performance Computing (HPC) . Any architecture of HPC would require scale up with time to keep up the performance. What's the best options and practices for this scale up? Can it scale vertically or horizontally? Is there any limitations around this or no limits i.e. infinite? The challenges on the way of AI are processing, data transport, data storage and retention and retrieval time . Yet there is no standard across this segment as this a new evolution and shall take time to establish. So there are possibilities of inefficiencies and cost difference across multitude of implementations happening in recent time...