Yuchen Jin· @Yuchenj_UW · X·· 4 小时前AI 评分62
AI 导读
作者以玩笑口吻称自己在 Databricks 的工作就是买大量 GPU 再读 DeepSeek 博客。其引用的内容提到 DeepSeek 正式开源大规模云端推理框架 DeepInfer，基于它为 DeepSeek V4.1 Flash 提供高吞吐低延迟服务，单集群吞吐 155 trillion tokens/天，按非高峰价格计算收入 $38 每卡每小时；引用者推算即便按 $7 每 GPU 小时租用 B300，毛利率约 82%，并称 DeepInfer 同时支持 NVIDIA Blackwell 和华为昇腾。
正文 · 原文
My entire job at Databricks:
Step 1: Buy a sh*tload of GPUs
Step 2: Read DeepSeek blogs
Step 3: Profit
“A single cluster currently serves 155 trillion tokens/day, generating $38 in revenue per GPU-hour (at off-peak pricing).” Huge flex from DeepSeek. Not only did they open-source their inference engine, but they also open-sourced their revenue numbers. They didn't disclose which GPUs they're using here, but DeepInfer supports both NVIDIA Blackwell and Huawei Ascend. Even assuming they're renting B300s at $7 per GPU-hour, that's ~82% gross margin.在 X 查看被引用的帖子
来源：Yuchen Jin · x.com