DeepSeek's escape from CUDA is unfinished: V3 trained on Nvidia with hand-tuned Nvidia instructions; V3.2 shipped kernels in both CUDA and TileLang; V4 achieved day-zero support on Huawei's Ascend chips, but DeepSeek's Tile kernels library still requires CUDA as a backend today.

factualpending

Speaker

Unidentified Speaker — DeepSeek's Invisible Founder Just Talked for 4 Hours. It Le… [BcKQk0NeBV0]

Evidence Quote

“V3 trained on Nvidia, down to hand-tuned Nvidia instructions. V3.2 shipped kernels in both CUDA and TileLang. V4 got day zero support on Huawei's Ascend chips, and DeepSeek's own Tile kernels library still requires CUDA today.”

Source

DeepSeek's Invisible Founder Just Talked for 4 Hours. It Leaked.— Squintist
Created: 8/12/2026, 6:15:41 PM

Your notes

Loading notes...