Vera Rubin 的容量與吞吐增益目前仍屬 NVIDIA 預測;相較之下,Lambda 數據是已公布的部署驗證結果。
What did Nvidia present on the opening day of its AI Infra Summit about transforming AI factories from systems optimized primarily for raw pNvidia’s AI Infra Summit focused on increasing useful AI output within fixed power budgets.
AI 提示詞
Create a landscape editorial hero image for this Studio Global article: What did Nvidia present on the opening day of its AI Infra Summit about transforming AI factories from systems optimized primarily for raw p. Article summary: Nvidia’s opening-day message was that an AI factory should be measured not only by peak compute, but by validated agentic tokens produced per megawatt—requiring integrated design across silicon, systems, networking, soft. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
openai.com
AI 基礎設施的瓶頸,愈來愈不只是機櫃裡能塞進多少加速器,而是資料中心究竟能取得多少電力。NVIDIA 在 2026 年 AI Infra Summit 首日提出的主張是:衡量 AI 工廠,關鍵不應只看峰值運算效能,而是看它在既定供電上限內,能產出多少有用的代理型 AI Token/每百萬瓦。 23
NVIDIA 超大規模與高效能運算副總裁 Ian Buck 在美國加州聖塔克拉拉的活動中闡述這項策略。其重點不只在 GPU,而是把運算、網路、推論軟體、散熱、電源控制與電網營運協同整合;當電力成為硬性限制時,AI 工廠必須靠整體系統設計維持產能。 123
在 NVIDIA 的 Eos AI 工廠中,Emerald AI 的 Conductor 平台與 Silicon Valley Power 合作參與商業化彈性負載計畫。NVIDIA 表示,該系統已回應超過 200 次公用事業需求訊號;在一次展示事件中,設施耗電量在不到一分鐘內由 4 MW 降至 3 MW。做法是降低可彈性調整的運算工作優先順序,同時讓優先等級較高的推論持續運作。 23