The possibility of a beryllium-steam reaction during severe accidents in the International Thermonuclear Experimental Reactor (ITER) is a safety concern because the hydrogen produced from this reaction could pose a flammability or detonation hazard. The physical mechanisms governing the production of hydrogen are examined, and the sequence of events during a postulated ex-vessel loss-of-coolant accident (LOCA) are presented. A MELCOR simulation of an ex-vessel LOCA with simultaneous failure of the plasma shutdown system indicates that an in-vessel breach of the coolant system occurs because of first-wall melt-through. For the ITER interim first-wall/shield-blanket (FW/SB) design, this accident results in ∼67 kg of hydrogen being produced. A similar simulation for the divertor predicts only 0.3 kg of hydrogen because of additional cooling experienced by the divertor during the blowdown of coolant into the vacuum vessel. There is evidence to indicate that beryllium evaporation from the first wall at a surface temperature of 1100°C is enough to cause plasma termination through beryllium evaporation. This plasma termination occurs prior to first-wall melt-through and could minimize or eliminate significant hydrogen production. Sensitivity studies were performed by varying the first-wall temperature at which plasma termination and in-vessel breach occurs for an ex-vessel LOCA scenario. This study shows that if the plasma is terminated before 150 s (i.e., a maximum first-wall temperature of 777°C) after the ex-vessel LOCA, the amount of hydrogen generated is ∼1 kg, which is well below the flammability limit of 10 kg and gives a reasonable margin for model uncertainty. Other sensitivity studies using the FW/SB model indicated a relatively weak dependence of the hydrogen produced on in-vessel and ex-vessel breach size. In addition, a 60% reduction in coolant inventory resulted in only a one-third decrease in hydrogen production from the base case. Preliminary calculations for an in-vessel LOCA indicate that 100 kg of 50-µm dust in the vacuum vessel could generate 2 kg of hydrogen.