Node-coincident nesting superimposes two lattice families on one periodic grid so that their struts merge at shared nodes into a single monolithic frame, printable under the same rules as its parents. This work quantifies the effect of such nesting on the initial compressive stiffness – the property that governs deflection-limited use of lattice cores – at equal nominal total relative density. An experimentally anchored modelling chain is built on compression tests of four stereolithography-printed families (Simple Cubic, Face-Centred Cubic, Diamond and Fluorite) at relative densities of 10% and 20%. Three model fidelities – a beam model, a node-compensated beam model and a full solid model – use a global effective modulus, a global effective diameter correction and topology-specific node-compensation parameters. The calibrated compensated beam gives an overall mean absolute percentage error of 16% across all eight experimental configurations and approximately 23% for the held-out subset (SC-10, FC-10 and FC-20); resolved-solid comparisons are reported where available. The calibrated chain then screens seventeen candidate nested architectures and one internal-consistency case drawn from an eight-family pool that includes two cells used exclusively as guests. Embedding a stretch-dominated guest in a bending-dominated host raises the stiffness by up to +117% at equal nominal total relative density, yet no union outperforms its better parent (−25% to −77%), in agreement with Gibson–Ashby scaling: nesting redistributes mass between load-transfer mechanisms but does not create a more efficient one. Nesting is therefore a reinforcement strategy for architecture-constrained hosts, and the experimentally anchored chain enables simulation-driven screening in which testing is reserved for selected candidates.