Precision medicine envisions a future of effective diagnosis, treatment, and prevention grounded in precise understandings of the genetic and environmental determinants of disease. Given that the original genome-wide association studies represented a predominately European White population, and that diversity in genomic studies must account for genetic variation both within and across racial categories, new research studies are at a heightened risk for inadequate representation. Currently biological samples are being made available for sequencing in biobanks across the USA, but the diversity of those samples is unknown. The aims of this study were to describe the types of recruitment and enrollment materials used by US biobanks and the diversity of the samples contained within their collection. Biobank websites and brochures were evaluated for reading level, health literacy, and factors known to encourage the recruitment of minorities, such as showing pictures of diverse populations. Biobank managers were surveyed by mail on the methods and materials used for enrollment, recruitment, consent, and the self-reported race/ethnicity of biobank participants. From 51 US biobanks (68% response rate), recruitment and enrollment materials were in English only, and most of the websites and brochures exceeded a fifth-grade reading level. When compared to the 2015 US Census, self-reported race/ethnicity of participants was not significantly different for Whites (61%) and blacks (13%). The percentages were significantly lower for Hispanics and Latinos (18 vs. 7%, p = 0.00) and Hawaiian/Pacific Islanders (0.2 vs. 0.01%; p = 0.01) and higher for Asians (13 vs. 5%, p = 0.01). Materials for recruitment predominantly in English may limit participation by underrepresented populations.