Samsung is expected to more than double output of its HBM4 and HBM4E DRAM

Samsung is preparing to more than double its HBM4 and HBM4E DRAM output next year, shifting monthly wafer capacity toward roughly 250,000 and moving about 80% of production to the newer HBM4 family. HBM, high-bandwidth memory, is the memory stacked directly onto AI accelerators like Nvidia's GPUs, and it has been one of the tightest links in the AI hardware supply chain over the past two years. A near-doubling of Samsung's output, combined with the shift toward the faster HBM4 generation, signals more accelerator memory reaching the market next year, which matters for anyone forecasting GPU availability, pricing, or lead times on AI infrastructure builds. It also indicates confidence from one of the three major HBM suppliers, alongside SK Hynix and Micron, that demand for next-generation accelerators will keep climbing rather than plateau.