ArXiv caps submissions at two per month as AI paper flood overwhelms the preprint server

Your submitted research papers will face longer review times on arXiv.

arXiv, the popular preprint server, is implementing a new submission limit of two papers per person per calendar month, effective October 1, 2026. This decision comes in response to an overwhelming surge in submissions, which have more than doubled in two years to over 40,000 per month, largely driven by AI-generated content and tools. The platform's volunteer moderators are struggling to cope with the volume, leading to delays for high-quality papers. The new cap applies across all categories and only to the submitting author, with co-authors unaffected. Rejected papers still count towards the limit, but deleted papers before publication will refund the submission slot. arXiv has already tightened its policies recently, including a ban for papers with unverified LLM outputs. This move occurs as arXiv transitions to an independent nonprofit, supported by recent $17.2 million in funding aimed at technical upgrades, particularly for handling AI-generated content.

The new submission cap signals a growing strain on academic infrastructure due to the rapid proliferation of AI-assisted research, forcing established platforms to implement drastic measures. This indicates a potential bottleneck for disseminating novel findings in fields heavily impacted by AI, affecting the pace at which new discoveries are shared and validated within the scientific community. The challenge of moderating an influx of AI-generated or AI-assisted content highlights the need for more sophisticated automated review processes and a re-evaluation of quality control mechanisms in academic publishing.

This change directly impacts researchers by potentially slowing down their publication timelines and may encourage a more curated approach to preprints. For AI developers, it underscores the unintended consequences of their tools on the academic ecosystem and could spur innovation in AI-powered moderation or submission management systems. The situation at arXiv mirrors similar pressures faced by AI conference organizers, suggesting a broader industry-wide challenge in managing the output of AI-driven research and development.