OpenAI, Anthropic and Google Unite to Stop Chinese AI Model Copying
Three fierce domestic rivals are sharing intelligence to clamp down on Chinese competitors distilling their AI models — the AI cold war just got a NATO.
The news. The narrative. The timeline.
Bureau: Beijing
Three fierce domestic rivals are sharing intelligence to clamp down on Chinese competitors distilling their AI models — the AI cold war just got a NATO.
DeepSeek's next-generation V4 model will run entirely on Huawei Ascend chips, marking a major milestone in China's push for AI hardware self-sufficiency.
More than 100 Baidu Apollo Go robotaxis stopped simultaneously in Wuhan on April 1 due to a system malfunction, stranding passengers and blocking roads.
Google released the Gemma 4 family under the Apache 2.0 license, offering its most capable open models yet with 256K context windows and commercially permissive terms.
Anthropic's Claude produced two working remote root exploits for CVE-2026-4747 in four hours -- the first time an AI has autonomously developed a full kernel-level RCE.
Anthropic paid over $400 million in stock for Coefficient Bio, a stealth biotech startup with fewer than 10 employees -- the highest per-head acquisition in AI history.
OpenAI's COO Brad Lightcap is shifting to 'special projects,' AGI development CEO Fidji Simo is on medical leave, and CMO Kate Rouch is stepping down -- all in one week.
Over 90,000 tech workers were laid off in the first 95 days of 2026, averaging nearly 973 cuts per day across the industry.