The paper presents a critical review of the relationship between ettringite formation and sulfate attack. Ettringite formation is associated with expansion. However, any ettringite-related expansion is not necessarily related to sulfate attack. Early ettringite formation (EEF) which occurs immediately (within hours) in a plastic fresh mixture does not produce any damaging expansion and is associated with the regulation of setting time of portland cement paste. Expansion after the hardening of cement paste can be advantageously used for development of chemical prestress in expansive cements. Delayed ettringite formation (DEF) occurs at late ages and the related heterogeneous expansion in a very rigid hardened concrete can produce cracking and spalling. Two different types of DEF are examined depending on the sulfate source: DEF caused by external sulfate attack (ESA) or internal sulfate attack (ISA). ESA, related to the interaction of environmental sulfate with the cement matrix, can be precluded by the use of impermeable concrete. On the other hand, ISA occurs in a sulfate-free environment due to the interaction of internal sulfate (from cement or gypsum contaminated aggregate) with calcium-aluminate hydrates of the cement paste. Two different mechanisms of DEF caused by ISA are examined. The first one is based on the thermal decomposition of ettringite in high-temperature cured concrete elements and the subsequent re-formation of ettringite at ambient temperature in a saturated atmosphere. According to the second mechanism ISA is based on a chain of three essential events (microcracking, late sulfate release, and exposure to water) and DEF could occur even at room temperature.