BREAKING
DeepSeek Building Own Inference AI Chip
The Plan So Far
1Hire chip engineers
2Talk to foundry & memory
3Inference-only chip
4Cut Nvidia reliance
0%
Nvidia premarket drop
0B
China AI chip market
0%
Huawei's share
H800 to Ascend 950
R1Jan 2025
Trained on Nvidia H800
Low-cost, high performance
V4Apr 2026
Optimized for Ascend 950
Huawei aided training
0$
per M input tokens
0M
token context
A Shift Toward Hardware Control
AI NEWS BLITZ
Reuters reports DeepSeek is developing its own AI chip for inference.