Abstract Background A vast number of organisms are known to produce structurally diversified cellulases capable of degrading cellulose, the most abundant biopolymer on earth. The generally accepted paradigm is that the carbohydrate-binding modules (CBMs) of cellulases are required for efficient saccharification of insoluble substrates. Based on sequence data, surprisingly more than 60% of the cellulases identified lack carbohydrate-binding modules or alternative protein structures linked to cellulases (dockerins). This finding poses the question about the role of the CBMs: why would most cellulases lack CBMs, if they are necessary for the efficient hydrolysis of cellulose? Results The advantage of CBMs, which increase the affinity of cellulases to substrates, was found to be diminished by reducing the amount of water in the hydrolytic system, which increases the probability of enzyme-substrate interaction. At low substrate concentration (1% w/w), CBMs were found to be more important in the catalytic performance of the cellobiohydrolases TrCel7A and TrCel6A of Trichoderma reesei as compared to that of the endoglucanases TrCel5A and TrCel7B. Increasing the substrate concentration while maintaining the enzyme-to-substrate ratio enhanced adsorption of TrCel7A, independent of the presence of the CBM. At 20% (w/w) substrate concentration, the hydrolytic performance of cellulases without CBMs caught up with that of cellulases with CBMs. This phenomenon was more noticeable on the lignin-containing pretreated wheat straw as compared to the cellulosic Avicel, presumably due to unproductive adsorption of enzymes to lignin. Conclusions Here we propose that the water content in the natural environments of carbohydrate-degrading organisms might have led to the evolution of various substrate-binding structures. In addition, some well recognized problems of economical saccharification such as unproductive binding of cellulases, which reduces the hydrolysis rate and prevents recycling of enzymes, could be partially overcome by omitting CBMs. This finding could help solve bottlenecks of enzymatic hydrolysis of lignocelluloses and speed up commercialization of second generation bioethanol.