Cytomorphic circuits on VLSI chips, which are founded on common Boltzmann thermodynamic laws between electronic transistor flow and chemical reaction flux, can rapidly simulate arbitrary deterministic and stochastic biological pathways. Circuit design and simulation are therefore useful for quantitative drug-cocktail formulation for multiple diseases, including those involving complex viral-immune interactions. Due to the presence of multiplicatively nonlinear positive and negative feedback, even simplified viral-immune interactions can lead to highly unpredictable behavior. Here, we use prior biological data to parameterize and simulate a simplified viral-immune cytomorphic circuit, whose equations correspond to those of a Lorenzian chaotic system. Cytomorphic chip data exhibit highly unpredictable behavior for certain ranges of parameters. Thus, as the recent pandemic shows, life-and-death outcomes to the same drug or drug cocktail often are highly variable or highly sensitive to patient-specific parameters. Cytomorphic circuits may be useful for designing, predicting, gaining insight into, or reducing therapeutic variability and sensitivity in the future, especially in complex disease models that are necessarily more empirical.