DeepSeek's next-generation V4 model will run on Huawei chips

DeepSeek 下一代模型 V4 将跑在华为芯片上。

DeepSeek delayed V4 for months and rewrote some low-level modules with Huawei and Cambricon so it runs on Huawei's Ascend 950PR, with launch expected in weeks, per The Information. The post cites 112GB memory, 1.4TB/s bandwidth, 600W power, and FP4 inference support; it does not disclose V4 size, pricing, or measured performance.

Why it matters: This clears HKR-H/K/R: Huawei-chip deployment is a strong hook, the report includes concrete module and chip details, and the China compute-stack angle will travel. It stays below 85 because this is pre-release reporting; model size, price, and real benchmarks are undisclosed.