Running DeepSeek V4 on GB300 NVL72
DeepSeek V4 has 1600 billion parameters. GB300 NVL72 offers 20,700 GB of HBM3e across a 72-GPU rack memory. This page puts the two side by side.
How much memory does DeepSeek V4 need?
Weights take the parameter count multiplied by the format size, plus a 15% runtime margin for activations and buffers. The attention cache sits on top: it grows with context length and with the number of concurrent requests, so it is sized case by case.
| Format | Weights in memory | On this machine | What the format costs |
|---|---|---|---|
| FP16 | 3,680 GB | fits | full precision, the quality reference |
| FP8 | 1,840 GB | fits | negligible loss on most tasks |
| NVFP4 | 920 GB | fits | Blackwell format, native FP4 compute |
| Q4 | 920 GB | fits | block quantisation, broadly supported |
What is left to serve requests?
DeepSeek V4 fits comfortably on GB300 NVL72. In FP16, its weights take about 3,680 GB of the 20,700 GB available, leaving 17,020 GB for the attention cache and concurrency. The remaining margin decides how many concurrent requests and how much context the machine sustains. A margin below a quarter of memory forces a limit on context or on concurrency.
GB300 NVL72 ranges from from 3,9 M€, indicative and not contractual. The machine targets the GE segment.
Where else can DeepSeek V4 run?
See the DeepSeek V4 and GB300 NVL72 fact sheets.
How much memory does DeepSeek V4 need?
In FP16, weights take about 3,680 GB including the runtime margin. The attention cache sits on top and depends on context.
Does DeepSeek V4 fit on GB300 NVL72?
DeepSeek V4 fits comfortably on GB300 NVL72. In FP16, its weights take about 3,680 GB of the 20,700 GB available, leaving 17,020 GB for the attention cache and concurrency.
Method
Parameter counts and memory figures come from the site fact sheets. Memory footprints are calculated, not measured: a reading on real hardware may differ depending on the engine and the exact weight format. Prices are indicative and not contractual.