Dylan Patel 认为谷歌 TPUv7 在智能体推理的每美元性能上已追平或超越 NVIDIA,迫使 NVIDIA 紧急调集工程力量应对。
TPU 在每美元性能上更优或具备竞争力,对 NVIDIA 高层来说是一种不愿接受的新体验。TPU 在 Qwen3.5 397B FP8 上发布首个 AgentX 结果、显示 TPUv7 FP8 甚至能击败 GB300 NVL72 FP4 后,NVIDIA 立即派出由工程师和智能体集群组成的突击队,在中美两地加班到深夜和周末。
原文 (en)
TPUs delivering better performance per dollar or competitive performance per dollar is a new experience that NVIDIA senior leadership is unwilling to accept. As soon as TPU posted its first AgentX result on Qwen3.5 397B FP8 showing that even TPUv7 FP8 could beat GB300 NVL72 FP4, NVIDIA tasked a strike force of human engineers and agent swarms to work 997, late into the night until 2am local time and on weekends, across USA and mainland China.