Array genomic hybridization (AGH) can detect chromosomal gains or losses that are 100 times smaller than those identifiable by conventional cytogenetic methods. Genome-wide AGH can identify genomic imbalance that causes birth defects and mental retardation at least twice as frequently as conventional cytogenetic analysis. Using AGH as a prenatal test for fetal genomic imbalance offers the promise of detecting pathogenic gain or loss of genomic material more quickly and much more frequently than current methods. However, the chance of finding a result of uncertain clinical significance is much greater than with conventional cytogenetic analysis, and the benefit–cost ratio of doing AGH in addition to conventional cytogenetic analysis in pregnancies at high risk for Down syndrome is likely to be poor. Very little is known about the natural history and range of clinical variability associated with most pathogenic submicroscopic copy number variants (CNVs). It seems doubtful that patients can be adequately counseled for prenatal AGH testing in most cases because the risks and benefits are unknown. At present, AGH should be offered for prenatal diagnosis only if the pregnancy is at especially high risk of having a pathogenic CNV or if AGH is being done as part of a clinical trial. Copyright © 2008 John Wiley & Sons, Ltd.