Several published studies claim that high rates of N2 fixation occur in sugarcane and sorghum, and have ascribed this result to infection by the bacterium Gluconacetobacter diazotrophicus, abetted by arbuscular mycorrhizal infection (Glomus clarum). These results have not been confirmed within Australia. In this study, G. diazotrophicus was detected in stalks of field-grown sugarcane in Australia (based on phenotypic tests, and a PCR test using species-specific primers developed to amplify a fragment of the G. diazotrophicus 16S rRNA gene). Isolates were nitrogenase positive (acetylene reduction assay) in vitro. However, in glasshouse trials involving inoculation of sugarcane setts with G. diazotrophicus, co-inoculation with mycorrhizae, and plant growth under low N status, recovery of bacteria from maturing plants was variable. At 165 days from planting, no appreciable N2 fixation, as assessed by dry weight increment, N budget, or 15N ratio, of either an Australian or a Brazilian cultivar of sugarcane, or a sorghum cultivar, was achieved. We conclude that a N2-fixing sugarcane–G. diazotrophicus association is not easily achievable, being primarily limited by a lack of infection.