NVIDIA 携手 Palantir 用 Nemotron 和 Foundry 编码供应链专业知识
原标题：From Wafer-Out to First Token: Codifying Supply Chain Expertise with Nemotron and Palantir Foundry
AI 摘要
NVIDIA 与 Palantir 合作，利用 Palantir Foundry 的 Ontology 构建数字供应链智能指挥中心，将材料、工厂、承诺、产能等数据统一为受治理的数据层。NVIDIA 使用开源 GPU 加速优化库 cuOpt 将关键材料分配问题建模为混合整数线性规划，以最小化所有权时间（TOO）为目标，并反馈约束瓶颈。双方通过回测历史分配决策发现，规划者的隐性知识（如邮件、天气、地缘政治事件）未被求解器捕捉，因此需要将人类专业知识编码化，形成持续积累的 AI 飞轮。
正文节选
NVIDIA has one of the largest and most complex supply chains in the world, and its performance is measured from wafer-out to first token. The interval is in two parts. Time-to-rack runs from silicon leaving the fab to an assembled system arriving on a data center floor. Time-to-token covers everything thereafter: power, cooling, networking, and the software stack that makes the infrastructure productive on day one. NVIDIA Grace Blackwell NVL72 platforms draw on millions of parts and thousands of