As foundation grantmaking has grown over the last decade, scrutiny from potential regulators and the general public has increased. However, little is known about foundations' decision-making processes for grant awards. The unique status of American Indian and Alaska Native tribal governments, their increased capacity over the last 30 years, and the disproportionate socioeconomic disadvantage of Native peoples' position Native communities as potentially significant foundation grantees. Yet research shows that foundation's grantmaking to Native causes generally, and to tribal governments in particular, is disproportionately low (Hicks and Jorgensen, 2004, 2005). This study applies a principal-agent framework to the organizational problem of large foundations' grantmaking to Native causes. Collectively, foundations are challenged to find the best-positioned grantees to fulfill their mission and create significant impact with their investments. They have the choice of agent types, who vary by level of risk, to whom they award grants: intermediary organizations (the safest choice), nonprofit organizations (the referent choice), and tribal governments (the riskiest choice). Foundations use the information that they have available about potential recipients, including information about recipient types, to make grantmaking decisions that minimize the risk that the grantees will act opportunistically or contrary to the foundation's intent. This study applied qualitative and quantitative methods to understand and explain large foundations' grantmaking to Native America. Considerable differences in information sources and grant award decision-making processes were identified. Foundation representatives expressed substantial variation in the factors their foundations considered in Native grant award decisions and the weight they attributed to relative factors. Relationships---the use of colleagues and networks---seemed paramount in assessing potential recipient organization's missions, positioning on an issue, community relationships, and capacity. The multinomial logistic regression model examined tribal governments relative to intermediary organizations (model one) and nonprofit organizations relative to intermediary organizations (model two). While the intercept for model one was significant (p < .05), the intercept for model two was not. Foundation type, total foundation grantmaking, and peer tribal government grantmaking contribute to predicting large foundations' choice of intermediary organizations relative to tribal governments. The geographic scope of foundations' grantmaking is a meaningful constraint on choice of recipient type.