We investigated the association between 2009 IOM recommendations and adverse infant outcomes by maternal prepregnancy body mass index (BMI). Birth outcomes for 570,672 women aged 18-40 years with a singleton full-term live-birth were assessed using 2004-2007 Florida live-birth certificates. Outcomes included large-for-gestational-age (LGA) and small-for-gestational-age (SGA). Associations between gestational weight change and outcomes were assessed for 10 BMI groups by calculating proportions, and logistic regression modeling was used to produce adjusted odds ratios (aORs) to account for the effect of confounders. We created comparison categories below and above recommendations using 2009 IOM recommendations as a reference. Of importance, 41.6% of women began pregnancy as overweight and obese and 51.2% gained weight excessively during pregnancy on the basis of 2009 IOM recommendations. Proportions of LGA were higher among obese women and increased with higher weight gain. Compared with recommended weight gain, aORs for LGA were lower with less than recommended gain (aOR range: 0.27-0.77) and higher with more than recommended gain (aOR range: 1.27-5.99). However, SGA was less prevalent among obese women, and the proportion of SGA by BMI was similar with higher weight gain. Gain less than recommended was associated with increased odds of SGA (aOR range: 1.11-2.97), and gain greater than recommended was associated with decreased odds of SGA (aOR range: 0.38-0.83). Gestational weight gain influenced the risk for LGA and SGA in opposite directions. Minimal weight gain or weight loss lowered risk for LGA among obese women. Compared with 1990 IOM recommendations, 2009 recommendations include weight gain ranges that are associated with lower risk of LGA and higher risk of SGA. Awareness of these tradeoffs may assist with clinical implementation of the 2009 IOM gestational weight gain recommendations. However, our results did not consider other maternal and infant outcomes related to gestational weight gain; therefore, the findings should be interpreted with caution.