To improve grid reliability, resilience and cost performance, energy companies increasingly rely on AI and other data science techniques to process large amounts of aggregated data. Typically, aggregation means centralizing customer data to train algorithms at significant time and cost while introducing security and privacy risks. KnowlEDGE investigates the advantages of using a distributed (federated) analysis strategy to se-cure value from industrial, commercial and residential smart meter data in multiple locations, in particular targeting DSO use cases, and will in-vestigate the feasibility of conducting analysis of data at the grid edge. KnowlEDGE includes the development of federated analytics algorithms, application to smart meter data, and testing on meter infrastructure in the laboratory and in the field. The project will establish that accessing / analysing data in its source locations at the grid edge provides time, cost, security and privacy benefits using a phased series of proof of grid-edge concept implementations.