Donation-based crowdfunding platforms often claim to pursue equitable outcomes for all beneficiaries, yet many face criticism for failing to do so across different demographic profiles. In response, platform managers are eager to understand how these inequities emerge and explore solutions to address them. In this research, we show that the degree of facial emotional expressiveness of beneficiaries in uploaded images can differentially impact donation amounts for White vs Black beneficiaries. Drawing on social vision theory, we propose that facial emotional expressiveness in images combined with the race of the faces activates racial stereotypes of emotion expression that result in differential donation amounts to Black and White individuals. Analyzing a sample of 4,153 campaigns from GoFundMe between June 2021 and September 2022, along with a follow-up experiment, we find that higher facial emotional expressiveness is associated with significantly lower donation amounts for Black compared to White beneficiaries. Further exploring our moderating constructs reveals that the use of call-to-action cues, affective messaging, and race-gender homophily cues can attenuate the activation of stereotypes and therefore reduce differences in donation amounts between racial groups. Based on these findings, we offer targeted recommendations for platform managers to help reduce racial inequities in crowdfunding outcomes.