arXiv, the open-access preprint repository that has become essential infrastructure for AI/ML and scientific research publishing, is separating from Cornell University to form an independent nonprofit organization. As part of the transition, arXiv is recruiting a CEO at a $300,000 annual salary — a notable professionalizing step for a platform that began in 1991 as a collection of shell scripts on a NeXT machine, created by physicist Paul Ginsparg at Los Alamos. Cornell has hosted arXiv since 2001, and the spinout marks the end of a quarter-century of university stewardship. The news was flagged by mathematician John Carlos Baez on Mathstodon.
The organizational shift is running in parallel with an infrastructure change: arXiv has migrated its servers from Cornell's on-premises systems to Google Cloud. Google is already a Gold Sponsor of arXiv, and the cloud migration — which traces its planning roots to early 2023 — reduces arXiv's dependence on Cornell's physical plant while deepening its relationship with a corporate backer. A well-paid executive at the helm and a cloud contract with a major sponsor are not, on their own, alarming signs. But they do mark a clean break from the informal academic model that defined arXiv for three decades.
The governance questions that follow are real, particularly for the AI/ML field, where arXiv preprints are the primary vehicle for rapid knowledge sharing. On Mathstodon, Baez noted the irony of a repository built on open-science principles growing reliant on corporate infrastructure, and asked whether a more distributed model — accountable to the academic libraries that actually use arXiv — might be worth pursuing. That question does not have a settled answer. The board arXiv assembles, and the CEO it hires, will go a long way toward answering it.