AI 导读
据 The Information 报道,梁文锋向投资人表示 DeepSeek 计划使用华为芯片训练,950DT 可能于 2026 年 Q4 或 2027 年 Q1 到货。他 reportedly 正在训练 2T 参数模型并规划 8T 模型,高于 V4 的 1.4T;其估算类 GPT 模型需 5 万块 Nvidia GB300 或 20 万块华为 950DT。由于华为芯片与 HBM 供应有限,DeepSeek 仍需更多 Nvidia B 系列算力。
正文
According to The Information, Liang Wenfeng told investors DeepSeek plans to use Huawei chips for training, potentially 950DTs arriving in Q4 2026 or Q1 2027.
He is reportedly training a 2T-parameter model and planning an 8T model, up from V4’s 1.4T.
Liang estimated a GPT-like model could require 50,000 Nvidia GB300s or 200,000 Huawei 950DTs. Limited Huawei chip and HBM supply means DeepSeek still needs more Nvidia B-series capacity.
来源:@thexpin · x.com