Norway’s 2 petabytes of Huawei flash storage and LLM training
Summary
Norway’s National Library is developing a sovereign Norwegian-language LLM and leveraging 2 PB Huawei OceanStor Dorado flash storage in its AI training pipeline. The project involves in-house compute (Nvidia DGX H200), a two-stage data pipeline, and the Sigma2 Olivia supercomputer; governance, evaluation, and orchestration remain ongoing challenges. A takeaway is Huawei storage's critical role and the broader implications for sovereign, language-specific AI models.