The problem of multiple-input–multiple-output (MIMO) feasiblity refers to whether it is possible to support specified numbers of streams allocated to the links of an MIMO network while canceling all interference. In unilateral interference cancellation, nodes account only for interfering links that they have been assigned to cancel and ignore other interfering links. We present several different formulations of the unilateral MIMO feasibility problem and use these formulations to analyze the problem's complexity and develop heuristic feasibility algorithms. We first prove that the general unilateral feasibility problem is NP-complete. We then identify several special cases where the problem is solvable in polynomial time. These include when only receiver-side interference cancellation is performed, when all nodes have two antenna elements, and when the maximum degree of the network's interference graph is two. Finally, we present several heuristic feasibility algorithms derived from different problem formulations and evaluate their accuracies on randomly generated MIMO networks.