The Debate About Chinese Open-Source AI
Axios, meanwhile, reported on Monday that last week’s release of Kimi 3 from Chinese firm Moonshot had stoked conversations about banning foreign open source AI (more on the new Kimi model). A ban would set off an uproar. More and more companies are finding that ever-advancing Chinese models—such as the one from Moonshot released last week—are a good way to control their AI spending.
Chinese models have accounted for 30% of tokens used by U.S. firms since February, according to a report from Wall Street firm William Blair published on Monday, citing OpenRouter data. And in some cases, companies have found Chinese models can do things the top U.S. models can’t do, thanks to guardrails imposed on U.S. technology, as former White House AI czar David Sacks noted in an X post Sunday night.
In a separate X post, Sacks said the guardrails put on American models are “only making ourselves less competitive.” On Monday, investors Chamath Palihapitiya and Bill Gurley both warned of the damage to U.S. companies of banning Chinese open-source AI. As Palihapitiya put it, forcing U.S. companies to spend “50-100X more than their competitors abroad” on American tech will make for “financially impaired” American companies, which will eventually lead to problems for the AI labs.
Hopefully Trump can be persuaded not to go down this route—although given his stance on tariffs, never say never.