This article, based on dissipativity theory, aims to tackle the consensus tracking issue for Lipschitz nonlinear singular multiagent systems (MASs) with switching topologies and communication delays. Rooted at the leader node, a directed spanning tree is assumed to be contained in the union of all possible interaction graphs. Within the framework of topology switching controlled by a Markov chain, communication delays encountered in the data transmission process are reasonably considered to be time-varying and dependent on Markovian jump modes. By using tools from the stochastic Lyapunov functional technique, algebraic graph theory, and strict (Q,S,R)-α-dissipativity analysis, the consensus controller collecting delayed in-neighboring agents' information is designed to ensure stochastic admissibility and strict dissipativity of the resulting consensus error system. The theoretical analysis is validated by numerical simulations.