NVIDIA Slashes DeepSeek-V4 Pro Startup Time to Under 2 Minutes

NVIDIA announced that its ModelExpress technology reduced the startup time of the DeepSeek-V4 Pro model from 8 minutes to under 2 minutes. This optimization was achieved by transferring model weights directly into VRAM using a GPU-to-GPU RDMA path.

2026-07-25 ~ 2026-07-25 · 4 related posts

• NVIDIA says ModelExpress cuts DeepSeek-V4 Pro startup from 8 minutes to under 2 — NVIDIAAI · 2026-07-25

3 near-duplicate retellings: NVIDIAAI · RisingSayak · pmttyji