Speculative realists claim that almost all post-Kantian philosophy suffers from what they call correlationism. Correlationism, they claim, originated with Kant's response to Hume. However, one of the major figures associated with the speculative realist movement, Levi Bryant, has argued that Gilles Deleuze's transcendental empiricism is able to confront the challenge of correlationism. My central claim is that while Bryant is correct not to label Deleuze as a correlationist, his analysis does not go far enough in that it takes speculative realism on its own terms and does not move beyond the logic of correlationism. Drawing from the recent literature on Deleuze's transcendental empiricism and Deleuze's first book on Hume, I argue that Deleuze does not read Hume to produce a non-correlationist ontology, but rather develops a non-ontological constructivist philosophy. Transcendental empiricism, as he later comes to call it, conceives of philosophy in such a way that not only avoids the problem of correlationsism, but reveals that the speculative realist project itself fails to move beyond the logic of correlationism.