Open Access, as a concept and as a movement, has come a long way – we have seen paywalls dismantled, licences expanded, large amounts of research now freely available to anyone with an internet connection. But often, what we mean by ‘open access’ seems to be taken for granted: open access to what, exactly? And open on whose terms? For all its grand ambitions, the dominant model of open access has largely reproduced the epistemic hierarchies and structures that determine what counts as knowledge in the first place. The scholarship liberated under OA models remains overwhelmingly text-based, peer-reviewed under Western academic workflows and norms, written in English, and rooted in Euro-American intellectual traditions that have systematically marginalised Indigenous knowledges, Global South scholarship, oral and community-based knowledge practices, and ways of knowing that do not fit neatly into the journal article or monograph format. Open access was meant to be transformative, but for it to genuinely become so, we need to move beyond thinking of ‘open’ merely in financial terms, and start considering questions of form, format, norms, and legitimacy. Who gets to decide what counts as legitimate, scholarly, or rigorous research? What assumptions are we making about our validation and quality assurance processes? Who is excluded by these processes and assumptions? Our high-minded ambitions require more of us than just open licences; we must actively support open epistemologies, as well. Engaging with these questions may be both uncomfortable and illuminating for academic libraries. Our institutions are not neutral; they are created within, informed by, and help to reproduce many of the same Euro-American hierarchies we once believe openness would challenge and dismantle. Is it even possible for us to expand our conceptions of what and whose knowledge should be collected, preserved, surfaced, and legitimised, within the confines of our institutional structures? This talk does not pretend to have all the answers, but it does think the questions are ones that libraries can no longer afford to ignore.