Chinese AI gains split Silicon Valley over curbs on open models
Why it matters: The fight could shape US rules on AI exports, intellectual property and how developers compete with China.
Chinese advances in open-source AI have sharpened a split across Silicon Valley over whether Washington should restrict Chinese access to American models and data. Companies including OpenAI and Anthropic argue powerful AI systems should stay tightly controlled and have pressed regulators over Chinese rivals they accuse of improperly harvesting data. Microsoft, Nvidia, Meta, Palantir and IBM have backed open-source development, saying broader access speeds innovation and lets more people test security. The clash intensified after Chinese startups Z.ai and Moonshot AI released models that US outlets described as rivaling American labs, pulling the debate beyond boardrooms and into the Trump administration.