Schizophrenic, severe major depressive and bipolar disorders often do not respond sufficiently to the usual pharmacotherapy and therefore constitute the world's greatest public health problem (1). Thus, we have every reason to be happy that we can offer cognitive behavior therapy (CBT) as an effective non-pharmacological adjunct to the millions of people who are treated for those conditions worldwide. Obviously there are methodological problems as well as discrepant and negative results in numerous studies and meta-analyses on the effectiveness of complementary CBT in the above disease groups. Overall, however, the evidence speaks much more for than against the assumption that CBT or CBT-oriented strategies may improve the outcome of pharmacological treatment and prophylaxis, especially for treatment-resistant or chronic patients. On the other hand, the “promise of CBT” should not be overestimated in this important area of therapeutics. Taken as a whole, the hitherto achieved effects are only moderate, and data concerning bipolar disorders are not sufficient. Also, the evidence of efficacy for CBT should not mean that other approaches are now excluded from the search for useful psychotherapies for severe mental disorders. The traditional vulnerability-stress-coping models and also the newer salience theory of psychoses (2) offer a good framework for the development and use of psychological treatment strategies. They show, in fact, that different environmental stressors as well as personal and environmental protective factors may be involved in each individual case. CBT programs can target various but obviously not all of these psychosocial factors. Furthermore, important psychosocial influences such as trauma or migration interact with genetic-epigenetic and other biological factors (3). Therefore, even enthusiastic psychotherapists should not ignore that treatment and prophylaxis of severe mental disorders primarily deal with the current correction and future prevention of disorder-specific neuronal network pathologies in the human brain. The more we understand these pathologies, the more it will be possible to correct the network alterations by means of targeted neuromodulation. Research is currently testing new psychotropic drugs, neuroprotective substances and various brain stimulation methods to achieve this goal. The debate on “promise and limitations of CBT” reflects the presently unsatisfactory state of development. To date there are still no really causally effective psychotropic drugs available for the treatment of severe mental disorders, due to the lack of knowledge about etiology. That is why we try to combine substances which only influence the final pathogenetic pathways, and therefore are only partially effective, with complementary psychological interventions. Here, CBT has proven particularly useful, as it has a favorable impact on some of the psychosocial factors which are important for the release of and coping with the clinical symptoms. Nevertheless, the therapeutic problems are far from being solved, and the search for causal treatment approaches must go on. The directions for future research should include further testing of CBT in the initial high-risk states of psychosis or bipolar disorders (4). In persons who have already suffered from long courses of these diseases, with relapses or chronicity, CBT can only be an adjunct and at best produce the effects described in Thase et al's paper (5). Often the primary therapeutic goal of recovery is not at all achievable and what can be obtained is only coping with symptoms or loss of functions in order to reduce the burden of the disease and improve the quality of life. On the contrary, in the high-risk states, which can nowadays be characterized very well for psychoses (6) and increasingly better for severe mood disorders (7), we just observe initial changes of experience which, depending on the individual constellation of stressors and protective factors, can progress or not to the actual disease. The prevention of the outbreak of severe mental disorders is what is meant by the concept of indicated prevention (8). Right from the beginning, CBT was involved in the studies of indicated prevention of schizophrenia and has repeatedly been shown to be effective as the sole therapy without concurrent use of medication (9). It seems to be able to even prevent the transition of early cognitive disorders into attenuated psychotic symptoms and their transition into full schizophrenic symptoms (10). If this approach of indicated prevention is successfully transferred also to severe mood disorders, it promises to be much more useful in the fight against the world's greatest public health problem than the late adjunctive strategy (11).