DeepSeekV4 1.6T Day 0 to Day 43 Performance Over Time - Huawei, GB300 NVL72, MI355X, B200
From the article
Day 0 Inference Performance, InferenceX, 100x performance improvement in 26 Days, Cost per Million Tokens, Huawei 950DT Inference Trace Analysis
Continue reading on SemiAnalysisYou might also wanna read
DeepSeek V4 — Open-Weight Frontier, Huawei Chips, and 93.5% on LiveCodeBench
chatforest.com·1mo ago
DeepSeek reportedly developing custom AI inference chip to reduce reliance on Nvidia and Huawei
DeepSeek, the AI company that gained prominence in early 2025, is reportedly developing its own custom AI inference chip to reduce reliance
DeepSeek Developing In-House AI Inference Chip to Reduce Reliance on Nvidia and Huawei
DeepSeek, the Chinese AI chatbot company, is developing its own in-house AI inference chip to reduce reliance on both US-based Nvidia and Ch
China's DeepSeek plans to develop its own AI inference chips to reduce reliance on Nvidia and Huawei
Chinese AI startup DeepSeek, known for developing competitive large language models, is planning to design its own data center chips for inf
arstechnica.com·1d agoDeepSeek is building its own AI chip to cut reliance on Nvidia and Huawei
tech-ish.com·1d ago
DeepSeek Developing Custom Inference Chip Amid China’s AI Hardware Push
Techstrong Semi·2d ago

Comments
Sign in to join the conversation.
No comments yet. Be the first.