China's open-weights AI strategy is winning

Our read
American AI is safety-washing itself into irrelevance. While US labs waste compute and engineering hours lobotomizing their models to avoid offending HR departments, China is shipping raw, cheap, high-performance open-weights models directly to global developers. You cannot win a global tech race by hiding your tech behind a corporate paywall and a lecture on ethics.
Key findings
The American Safety Bureaucracy: Advocating for closed-source, heavily regulated, and proprietary AI systems under the guise of national security and safety.
The Open Weight Realists: Recognizing that China's open-weights offensive is winning over global developers by offering raw, accessible utility without the corporate lecturing.
A massive surge in developer adoption of DeepSeek and other open-weights models, coupled with growing frustration in the Silicon Valley ecosystem over the cost and sanitization of US proprietary APIs.
What happened
China is aggressively exporting high-performance open-weights AI models like DeepSeek to global developers, while US labs remain locked in a defensive crouch of proprietary safety-washing.
The fight
- The American Safety Bureaucracy
Advocating for closed-source, heavily regulated, and proprietary AI systems under the guise of national security and safety.
- The Open Weight Realists
Recognizing that China's open-weights offensive is winning over global developers by offering raw, accessible utility without the corporate lecturing.
The brief
China is aggressively exporting high-performance open-weights AI models like DeepSeek to global developers, while US labs remain locked in a defensive crouch of proprietary safety-washing.
** The American Safety Bureaucracy say Advocating for closed-source, heavily regulated, and proprietary AI systems under the guise of national security and safety.
The Open Weight Realists say Recognizing that China's open-weights offensive is winning over global developers by offering raw, accessible utility without the corporate lecturing.
Why now. A massive surge in developer adoption of DeepSeek and other open-weights models, coupled with growing frustration in the Silicon Valley ecosystem over the cost and sanitization of US proprietary APIs.
