{"ID":6267182,"CreatedAt":"2026-07-10T01:11:38.759438437Z","UpdatedAt":"2026-07-13T22:39:15.707437423Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.08407","arxiv_id":"2607.08407","title":"Who Needs DRAM? We Have Fiber","abstract":"The rising pressure on DRAM availability and contract pricing reflects generative AI's massive high-performance memory requirements. This pressure is heavily compounded by hyperscale data center expansion, which now consumes a significant portion of global DRAM output. In this work, we propose a new architecture: Fiber Memory, which reimagines the role of optical fiber in a hyperscale data center, deploying it as an active, recirculating delay-line memory for immutable data, such as large language model (LLM) weights. We present a data-parallel optical broadcast delay-line memory architecture that accounts for fiber's physical realities. By incorporating space-division multiplexed multi-core fibers (MCFs), passive optical tap-and-amplify interfaces, co-packaged optics (CPO), and regional all-optical regeneration, our case study evaluation demonstrates that Fiber Memory can eliminate redundant weight storage across 10,000 AI accelerators and reduce weight-delivery energy by over 70% compared to traditional HBM3e configurations.","short_abstract":"The rising pressure on DRAM availability and contract pricing reflects generative AI's massive high-performance memory requirements. This pressure is heavily compounded by hyperscale data center expansion, which now consumes a significant portion of global DRAM output. In this work, we propose a new architecture: Fiber...","url_abs":"https://arxiv.org/abs/2607.08407","url_pdf":"https://arxiv.org/pdf/2607.08407v1","authors":"[\"Hannah Atmer\",\"Thiemo Voigt\",\"Yuan Yao\",\"Stefanos Kaxiras\"]","published":"2026-07-09T12:33:24Z","proceeding":"cs.AR","tasks":"[\"cs.AR\",\"cs.DC\",\"cs.ET\",\"cs.NI\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}