The cost of investing in grain storage capacity at country receival points depends on the technology used (level of capital intensity) as well as annual variaation in the volume of grain produced. Peak load pricing theory, which was originally formulated to examine investment problems for public utility industries who face a fluctuating and uncertain demand, was applied to the grain storage problem. A cost minimization investment model is outlined which examines the problem of choosing the optimal combination of technology types, and the total level of storage to meet a fluctuating demand for storage. An examination or the marginal costs of grain storage reveals that the demand for storage should generally be met by using permanent horizontal storage technology to satisfy most demand, while using bunker storage to satisfy the peak demand arising form bumper harvests. This is in contrast to the widespread use of vertical storage in some states, particularly South Australia. It is revealed that the capital costs of vertical storage are far too high to justify investment for long term storage of grain. An examination of actual investment levels in Western Australia revealed a generally high level of overinvestment with some sites having twice as much storage as they should have had. A second major finding was that many sites appeared too small to justify the high fixed costs associated with investment in permanent storage facilities.