Spatially programmable origami networks enable high-density mechanical computing for autonomous robotics.

Mechanical computing enables logic decision-making, allowing direct computational integration into robotics to enhance their autonomy in complex environments. However, current non-universal logic designs hinder reconfigurability in multifunctional mechanical computing systems. Complexity-multifunctionality trade-offs limit mechanical computing materials to single logical operations and low computational density. Here, we address these limitations using origami metamaterials with reconfigurable conductive networks, enabling high-density programmable logic via physical reorganization. By rotating intra-gate elements to modify AND/OR-based Boolean cascades, the design reduces gates by 46.7% compared to standard arrays, executing arithmetic and comparison operations efficiently. Shared tree-like cascades allow multiple functions with minimal redundant gates. The system via Rubik's Cube mechanics supports three-axis reconfiguration of Buffer/NOT elements, achieving reconfigurable full-adder/subtractor and computational densities up to 1728. Integrated robotics demonstrate autonomous right-angled and curved path planning through reprogrammable half-adder/subtractor logic. This framework provides a universal, scalable design-methodology for high-density mechanical computing, with implications for robotics and embodied intelligence.