[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-deepseek":3},{"tag":4,"articles":11},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":10},"70323e59-46bb-4b00-93c0-03d6211f3206","DeepSeek","deepseek",5,"DeepSeek 代表新一輪開源與低成本大模型競爭，牽動推論效率、KV cache、GPU 配置與雲端部署成本。從模型架構到 NVIDIA 供應鏈，它正在改寫 AI 基礎設施的估值與選型。","DeepSeek marks a shift in open-weight LLM competition, where inference efficiency, KV cache behavior, GPU memory, and cloud deployment cost matter as much as raw model size. It also feeds into broader questions about NVIDIA, hardware supply, and AI infrastructure pricing.",[12,21],{"id":13,"slug":14,"title":15,"summary":16,"category":17,"image_url":18,"cover_image":18,"language":19,"created_at":20},"0ca0aadf-acab-49d2-8e28-2cb93ec7252f","will-nvidia-stock-recover-after-deepseek-en","Will NVIDIA Stock Recover After DeepSeek?","DeepSeek shook AI spending assumptions, but NVIDIA still controls the GPU supply chain. Here’s what the numbers say about a rebound.","industry","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775161855104-ayxh.png","en","2026-04-02T20:30:36.581851+00:00",{"id":22,"slug":23,"title":24,"summary":25,"category":17,"image_url":26,"cover_image":26,"language":19,"created_at":27},"b6caa87d-6766-486b-b510-0b27c6222f8e","nvidia-b300-vs-h200-deepseek-perf-en","NVIDIA B300 vs H200: Specs and DeepSeek Perf","B300 packs 288GB HBM3e and up to 8TB\u002Fs bandwidth. Here’s how it compares with H200 for DeepSeek inference and cloud costs.","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775161683086-zo6y.png","2026-04-02T20:27:39.316088+00:00"]