The self-healing capacity of bone for the reconstruction of small fractures is widely recognized. However, large bone defects still pose a big challenge for orthopedic surgeons. Tissue engineering using porous scaffolds is a new trend to solve this problem and to speed the healing process. When seeded with human cells and bone growth stimulating molecules, they enhance bone healing. Such porous supports try to mimic the properties of trabecular bone. The needed structural properties are: a high and interconnected porosity, pore sizes in the range of 100–500 µm, mechanical properties comparable to trabecular bone and a microporous surface to allow further coating, cell attachment and seeding. Starting from these requirements, it is clear that a cellular material has the ideal architecture for such biomedical scaffold. Polymeric, ceramic or metallic foam structures can be obtained by different manufacturing routes. In our group we optimized three methods, all starting from a powder suspension. They are the PU template technique, gelcasting and the 3D fiber deposition (3DFD) method. Ti and Ti–6Al–4V scaffolds produced by gelcasting and 3DFD can mimic the structural and mechanical properties of the trabecular bone. Hydroxyapatite (HA) and mixtures of hydroxyapatite and beta tricalcium phosphate (β-TCP) were tested for their rate of dissolution in simulated body fluid (SBF) and for their strength as function of their composition. Recently, it was proven that magnesium is a biocompatible and bio-absorbable material. Therefore, as a metallic scaffold it can be expected that it can fulfill both functions of the biomedical scaffold: to be load-bearing and to be bio-absorbable. Nevertheless, the realization of this ideal scaffold needs much more fundamental research. Some preliminary synthesis tests for producing porous Mg-scaffolds are given in this contribution.