The power industry is changing rapidly. Utilities in deregulated markets are forced to focus on increased customer service on a very competitive market. In order to become successful the utilities have to utilize all their resources. One such resource is the huge amount of information in different computer systems. This information can be refined and combined to show patterns and support conclusions that could be used to provide better service to the customers, gain market shares and increase profit. The key is to store this information in a data warehouse that is flexible enough to provide powerful data-mining capabilities, and powerful enough to handle the big information pressure caused by the constant inflow of measured data from the power network. To use data warehousing in the power industry needs special considerations. To be used as a reliable historian for process data, there are high requirements on performance, availability and redundancy. At the same time it must have all the flexibility and openness required to let the user perform e.g. ad hoc queries and data mining. This paper discusses some of these considerations, and possible realizations.