Mark RadarMARK RADAR
About
中 EN
Sign in
Event File AI OpenAI Jensen Huang Open-Source AI Models

Trump Administration Debates Response to China’s Open-Weight AI

1 reports · First detected 2026-08-05 · Last active 2026-08-05

Chinese developers are gaining ground with open-weight AI systems that can be downloaded, modified and deployed at lower cost than proprietary services. Moonshot AI’s Kimi K3 has approached models from OpenAI and Anthropic in some capabilities, accelerating adoption among U.S. startups. The advance has forced Washington to weigh cybersecurity and intellectual-property concerns against the risk that restrictions could raise costs, curb innovation and weaken U.S. competitiveness.

Trump administration officials in July considered sanctions, trade blacklisting and limits on U.S. cloud providers serving Chinese AI labs. Treasury Secretary Scott Bessent said on July 21 that Washington would examine possible intellectual-property theft through model distillation. Opposition then intensified: nearly 200 startup founders urged the White House to preserve access, while Nvidia and other technology companies defended open models. By Aug. 4, the administration had told developers that open-weight systems would be excluded from planned voluntary safety testing, signaling a shift toward bolstering domestic competition.

All Coverage

1 original reports
TECHNEWS.TW 2026-08-05
如何因應中國開源 AI 模型，川普政府內部爭論

The Backstory

The history behind this event
Industry Pushes Back as U.S. Weighs Chinese Open-Weight AI Curbs2026-07-27 · 3 reports · similarity 0.80

Moonshot AI’s release of Kimi K3 on July 16, 2026, revived the U.S.-China AI race that intensified after DeepSeek’s breakthrough in 2025. The Chinese open-weight model is competitive with flagship systems from Anthropic and OpenAI, while charging $15 per million output tokens, less than a third of Anthropic Fable 5’s $50. Its combination of strong performance, lower prices and downloadable weights has turned a national-security debate into a test of competition in the AI market.

Axios reported on July 20 that the Trump administration was weighing federal procurement rules, Entity List pressure and security-liability requirements for U.S. companies hosting Chinese models. On July 24, Hugging Face, Meta, Microsoft, Mistral, Nvidia and Replit were among companies signing an open letter against broad, “premature restrictions.” The group said open weights preserve competition and transparency, and argued that defenders need access to models with capabilities comparable to those used by attackers to detect, simulate and remediate cyber threats.

Mark Radar｜MARK RADAR
Past Issues・by MarkReadFintech
All times are in Taipei time (GMT+8)