Thermal limits of viable ectotherm development are threshold‐like and near‐symmetrical around the temperature of optimal performance and usually well within the thermal tolerance range of adult physiological traits. A proximate model is proposed to show that the interaction between reversible temperature inactivation of cell cycle proteins and their regulation can explain (1) the symmetry and (2) threshold character of thermal limits of viable embryonal and larval development in ectotherms. It is suggested that temperature inactivation of regulatory proteins mimics the decrease in concentration resulting from gene dosage change and transcriptional regulation during the cell cycle. If certain regulatory proteins have equal probability to be active or inactive at a certain temperature, cell division and, consequently, development becomes blocked. Model predictions were tested by comparing thermal tolerance limits as observed in viability experiments with 14 developing insect species with the estimated temperatures at which a hypothetical rate‐determining developmental enzyme has an equal probability to be active or inactive. These ‘expected’ thermal limits were derived from the Sharpe‐Schoolfield equation which describes temperature–differentiation rate reaction norms. In 21 out of 23 comparisons ‘expected’ thermal limits agree closely with the observed thermal tolerance limits. The implications of the model for thermal tolerance, thermal adaptation, epidemiology and life‐history strategies are discussed.