Long-acting injectable medicines are commonly developed by selecting a depot platform, adjusting formulation variables, and then evaluating the resulting release and pharmacokinetic behavior. This forward sequence can produce technically sophisticated formulations while leaving the governing therapeutic question incompletely specified: what depot behavior is required to generate an intended exposure trajectory across realistic patients, administration conditions, and manufacturing variation? This article develops Exposure-Constrained Depot Inverse Design, a proposed theory for designing the depot backward from a multidimensional target exposure specification. The approach decomposes the desired concentration–time behavior into admissible drug-input functions; maps those functions to molecular, material, formulation, administration, and process variables; represents depot formation, transformation, local tissue response, and systemic uptake as coupled dynamic processes; applies feasibility and manufacturability constraints before candidate acceptance; and requires forward evaluation across virtual populations and perturbation scenarios. The central contribution is a shift from maximizing a single output, such as nominal duration or cumulative release, toward identifying a robust feasibility envelope in which exposure requirements, depot physics, physiological variability, analytical observability, and product-development constraints remain simultaneously compatible. The theory distinguishes systemic exposure from in vivo input, intrinsic material behavior from the in vitro measurement system, and computational prioritization from experimental confirmation. Its principal limitations are non-uniqueness of the inverse problem, incomplete observability of depot states, model-form and parameter uncertainty, platform dependence, and limited human injection-site evidence. The proposed architecture may support more traceable formulation hypotheses and more informative validation studies, but it does not constitute a validated predictive model, clinical dosing framework, manufacturing control strategy, or regulatory decision tool.