In green parts of the plant, during illumination ATP and NAD(P)H act as energy sources that are generated mainly in photosynthesis and respiration, whereas in darkness, glycolysis, respiration and the oxidative pentose‐phosphate pathway (OPP) generate the required energy forms. In non‐green parts, sugar oxidation in glycolysis, respiration and OPP are the only means of producing energy. For energy‐consuming reactions, the delivery of NADPH, NADH, reduced ferredoxin and ATP has to take place at the required rates and in the specific compartments, since the pool sizes of these energy carriers are rather limited and, in general, they are not directly transported across biomembranes. Indirect transport of reducing equivalents can be achieved by malateoxaloacetate shuttles, involving malate dehydrogenase (MDH) for the interconversion. Isoenzymes of MDH are present in each cellular compartment. Chloroplasts contain the redox‐controlled NADP‐MDH that is only active in the light. In addition, a plastid NAD‐MDH that is permanently active and is present in all plastid types has been found. Export of excess NAD(P)H through the malate valves will allow for the continued production of ATP (1) in photosynthesis, and (2) in oxidative phosphorylation. In the latter case, the coupled production of NADH is catalysed by the bispecific NAD(P)‐GAPDH (GapAB) in chloroplasts that is active with NAD even in darkness, or by the specific plastid NAD‐GAPDH (GapCp) in non‐green tissues. When plants are subjected to conditions such as high light, high CO2, NH4+ nutrition, cold stress, which require changed activities of the enzymes of the malate valves, changed expression levels of the MDH isoforms can be observed. In nodules, the induction of a nodule‐specific plastid NAD‐MDH indicates the changed requirements for energy supply during N2 fixation. Furthermore, the induction of glucose 6‐phosphate dehydrogenase isoforms by ammonium and of ferredoxin and ferredoxin‐NADP reductase by nitrate has been described. All these findings are in line with the assumption that a changed redox state caused by metabolic variability leads to the induction of enzymes involved in redox poise.