English medium instruction (EMI) has become increasingly popular around the globe. Different stakeholders, including policymakers, teachers, students, and parents, believe that EMI would give rise to benefits in both content subject and second language (L2) English learning. However, this dual goal is not always realised. It is, therefore, essential to examine what is happening in the classroom. This study delves into a critical component of classroom interaction—teacher feedback. One unit of six EMI biology lessons taught by an experienced teacher to a class of Grade 11 students (aged 17 years old) in Hong Kong was observed, videotaped, and analysed. Teacher feedback was coded into different categories based on existing taxonomies (recasts, explicit correction, clue, elicitation, repetition, and clarification requests). These categories were further divided into feedback targeting language or content. Results indicated that overall, there were many elicitations but few explicit corrections. When considering the language-content split, content feedback was four times more frequent than language feedback. The content feedback was primarily output-prompting (e.g., elicitation), whereas language feedback was mostly input-providing (e.g., recasts). This chapter ends with implications for further research and pedagogy.