Moonshot AI's Kimi K3 model sparks US‑China AI dispute

Moonshot AI launched the open‑weight large language model Kimi K3 on 16 July 2026. The 2.8‑trillion‑parameter system achieved top scores on several benchmark leaderboards, outperforming OpenAI’s GPT‑5.6 and Anthropic’s Claude Fable 5 in coding and agent tasks. Demand was so intense that the company halted new subscriptions within two days of release.

U.S. officials, including White House OSTP director Michael Kratsios, accused Moonshot of “large‑scale industrial distillation” that allegedly copied Anthropic’s Claude Fable 5 data, warning of possible sanctions. The Chinese side denied the claim, calling it politicised. The controversy triggered a sharp fall in global semiconductor equities, wiping roughly $3.3 trillion from market cap, and prompted U.S. tech leaders to weigh policy responses.

Microsoft is testing Kimi K3 for Azure‑Copilot, projecting potential annual savings of up to $600 million, while CEOs Satya Nadella and Jensen Huang publicly supported the continued use of open‑weight models. Elon Musk praised the model as “Impressive.” The episode highlights the accelerating competition between Chinese open‑weight AI and U.S. closed‑source offerings, with implications for trade policy, AI safety, and the economics of large‑scale model deployment.