Physix Frontier · News Briefing Card (enbrief · Nov 17, 2026)
Study: US and Chinese AI insiders both question 'AI race' narrative
KEY FACTS
- Two scholars published an article on November 16 comparing official and netizen discussions of DeepSeek in China and the US.
- The study argues that 'who is winning' is the wrong question, and that the race is not a fact as described.
- DeepSeek trained on downgraded Nvidia H800 chips, which the US side views as evidence of a regulatory loophole.
- Chinese state media said the model shattered the illusion of a chip blockade and emphasized independent innovation.
- Zhihu users pointed out that the model still relies on Nvidia's CUDA software ecosystem.
KEY DATA
nearly $600 billionNvidia single-day market cap loss
PHYSIX OBSERVATION
The 'AI race' is more a framework repeatedly invoked by policy and capital than a verifiable fact. The same model is told as two stories in China and the US—threat and breakthrough—with chips, compute, and open source all bound to political narratives. For users, what warrants caution is that race rhetoric is now underwriting export controls and massive compute investment, and technical discussion is easily displaced by stance-first thinking.
Source: enbrief original report ↗
Physix Frontier