ShapSum is a multi-document summarizer. Given a document, it constructs a sentence similarity graph using lexical and semantic cues (WordNet, POS). Then it ranks sentences based on Shapley values from cooperative game theory to capture influence/coverage. The top-k nodes act as summary sentences for the document. We devised a novel approach to treating the problem of text summarization as a case of information diffusion. We used Shapley Value, a solution concept from cooperative game theory, for summary extraction. Our model outperformed the state-of-the-art results on the DUC dataset.