Our research explores the growth of surface complexity through careful attention to program and performance criteria. As this complexity emerges, however, we are repeatedly confronted with the realization that its cost compromises any of its performance gains. While the aggregation of repeatable units with variation from one unit to the next is achievable at a low cost through subtractive fabrication technologies (CNC milling, laser cutting, waterjet cutting), it is more difficult to achieve through casting or forming technologies (concrete casting, injection molding, vacuum forming). This is because formwork is not adaptable. Once you produce a mold, typically at a high cost, that mold makes one component only. If you want variation, a new mold must be produced for each new component. With the projects Hexwall and VarVac Wall we put forward a simple question: Can an intelligent, adaptable vacuum-forming mold be developed that allows for difference from one component to another without the necessity for multiple molds? The research positions our design efforts strategically at the front end of the fabrication process. Our goal is to develop a malleable tool that allows for endless variation in a fabrication process where variation is typically impractical.