We present a scalable Bayesian model for low-rank factorization of massive tensors with binary observations. The proposed model has the fol-lowing key properties: (1) in contrast to the mod-els based on the logistic or probit likelihood, us-ing a zero-truncated Poisson likelihood for bi-nary data allows our model to scale up in the number of ones in the tensor, which is espe-cially appealing for massive but sparse binary tensors; (2) side-information in form of binary pairwise relationships (e.g., an adjacency net-work) between objects in any tensor mode can also be leveraged, which can be especially use-ful in “cold-start ” settings; and (3) the model ad-mits simple Bayesian inference via batch, as well as online MCMC; the latter allows scaling up even for dense binary data (i.e., when the num-ber of ones in the tensor/network is also mas-sive). In addition, non-negative factor matrices in our model provide easy interpretability, and the tensor rank can be inferred from the data. We evaluate our model on several large-scale real-world binary tensors, achieving excellent compu-tational scalability, and also demonstrate its use-fulness in leveraging side-information provided in form of mode-network(s).