Vat photopolymerization (VP) additive manufacturing relies on the local irradiation of a photocurable resin to produce layers of solid material, that are assembled layer-by-layer to produce 3 dimensional (3D) parts. Irradiation can be provided in a serialized (e.g. rastered laser) or parallelized (e.g. spatial light modulator) fashion. The latter approach, exemplified by liquid crystal display and digital light processing forms of VP has considerable promise due to high-throughput nature of the printing process, and rapid technological progress of the underlying light engines. To predict final part geometry and performance requires thorough characterization of the unreacted resin and the light engine characteristics. Validation of final part geometry also requires advanced sub-voxel scale characterization tools. In this talk, we will discuss the rigorous material, process and part characterization tools needed for generating high veracity VP data sets suitable for simulation. We will use the exemplary case of the 2022 NIST Additive Manufacturing Benchmark Series problem: “AMB2022-07: Vat Photopolymerization Measurements of Cure Depth and Print Fidelity Vs. Varied Exposure Duration, Photopattern Dimensions, and Resin Characteristics.” In this problem we consider how photomasks and resins can be controlled and characterized to predict the feature-size dependence of the working curve – a measure of cure depth versus light dosage. We will then expand the discussion to cover how in situ and ex situ measurements of the voxel scale cure behavior can be used to fundamentally understand the printing process, and through conditional generative adversarial network machine learning models, predict printing performance. Overall, high veracity data plays a critical role in the progress of VP additive manufacturing for critical applications ranging from consumer goods to automotive to bioprinting.