Metal–organic frameworks (MOFs) have enjoyed considerable interest due to their high internal surface areas as well as tunable pore geometry and chemistry. However, design of optimal MOFs is a great challenge due to the significant number of possible structures. In this work, we present a strategy to rapidly explore the frontiers of these high surface area materials. Here, organic ligands are abstracted by geometrical (alchemical) building blocks, and an optimization of their defining geometrical parameters is performed to identify shapes of ligands which maximize gravimetric surface area of the resulting MOFs. A strength of our approach is that the space of ligands to be explored can be rigorously bounded, allowing discovery of the optimum ligand shape within any criteria, conforming to synthetic requirements or arbitrary exploratory limits. By modifying these bounds, we can project to what extent achievable surface area increases when moving beyond the present limits of organic synthesis. Projecting optimal ligand shapes onto real chemical species, we achieve blueprints for MOFs of various topologies that are predicted to achieve up to 70% higher surface area than the current benchmark materials.