A data lake is a centralized repository that stores large volumes of raw structured, semi-structured, and unstructured data in its native format, enabling schema-on-read analytics and machine learning.