Mammalian cell cultures are the most appropriate host cells for recombinant DNA derived products if complex protein structures have to be synthesized in their native form. Due to their physiological behaviour they grow either adherent or in suspension. For the attachment of adherent cells, microcarriers or wire springs can be applied to increase the internal surface of the bioreactor. Both systems provide a simplified media exchange but, however, show some limitations in scale up. In contrast, suspension culture systems as homogeneous systems independent of any carrier have not shown any limitation in scale up. Because most cell lines which are of commercial interest grow in suspension, this technology is best advanced and used in batch and continuous mode. Although mammalian cell cultures are sensitive to hydrodynamic shear forces, technologies for deep tank production are developed which allow stirrer tip speed of up to 1.5 m s −I sufficient for oxygen uptake, suspension of cells and homogeneous supply with nutrients. For long term bioprocesses without selection pressure it has to be considered that transformed cell lines might show genetic instability due to their variations of chromosomes. In addition, sterile technology becomes an important factor in long term bioprocesses. The decision as to which cell culture system should be chosen, whether batch or continuous processes should be applied essentially is based on the capital investment, the amount of material to be produced, genetic stability of the production cell line, reliability of sterile technology and the flexibility required in the production plant. Under the assumption that 20 kg of a protein have to be produced per year and the same product concentrations in the harvest fluid are reached in the batch process and for instance in the chemostat, it can be considered that the capital investment for one 10,000 1 batch process and a 2 X 2,000 1 continuous process, necessary to produce the amount of material, is comparable. Risk of microbial contamination or technical failure can be considered to be fairly low in the batch process. The economic risk for long term bioprocess in the chemostat can be expected to be medium and high in the perfusion system which is in the large scale not technically fully satisfactory. In addition, due to the longer down time period after contaminations and the start up of the continuous process, the annual yield of the batch process can be considered to be higher. To match the capacity of the batch process at equal capital investment either the flow rate or the product concentration has to be increased and this is only possible in a perfusion system with cell retention. Due to the long bioprocess period the continuous process is less flexible and can only handle a low number of products per year. Compared to the batch process, as long as the formulated bulks derived from different harvests of one bioprocess run meet the specification, they can be combined with other formulated bulks of the same bioprocess run or from a different bioprocess run. The necessary time required for process development and validation of the continuous process will be prolonged in the continuous process compared to the batch process. To ensure constant product quality the Master Cell Bank and the Master Working Cell Bank have to be characterized according to their identity, purity, safety and their genetic stability and productivity within and beyond the fixed bioprocess period. However, in both cases, the bioprocess has to be validated for constant productivity, product quality and viable cell count from harvest to harvest and bioprocess run to bioprocess run. In conclusion, continuous processes only can compete with batch processes at identical capital investments if higher cell densities and product concentrations in the harvest fluid are achieved. This will be possible if perfusion technology can be further developed to larger scales.