Via gizmodo.com
The Trump-linked DeFi protocol says WorldClaw operates independently, but a 38% ownership stake and shared stablecoin revenue tell a more complicated story.
World Liberty Financial, the DeFi protocol backed by the Trump family, is drawing a line between itself and WorldClaw, an AI model aggregator that has come under fire for offering access to dozens of models built by Chinese developers flagged for national security concerns. The line, it turns out, is more of a dotted one.
WLFI’s representative David Wachsman stated that WorldClaw is not managed by World Liberty Financial and operates independently. That framing, however, sits awkwardly alongside the fact that WLFI holds a 38% stake in WorldClaw and collects revenue every time someone pays for AI model access using the USD1 stablecoin on the platform.
The Reuters investigation that started the conversation #
The clarification came after a Reuters investigation spotlighted WorldClaw’s model library, which includes over 90 AI models. Of those, 43 were developed by companies that have been flagged by the US government for national security concerns, including firms linked to Alibaba and Baidu.
WorldClaw, founded in early 2026, bills itself as an “operating system for AI agents,” aggregating access to more than 300 AI models at competitive pricing.
The USD1 connection #
WorldClaw’s payment infrastructure runs on USD1, the stablecoin issued by World Liberty Financial. Every transaction on the platform generates revenue that flows back to WLFI, and by extension, to the Trump family, which holds a significant stake in the protocol.
Donald Trump Jr. and Eric Trump both promoted WorldClaw on social media prior to the controversy. WLFI has also acknowledged that Ryan Fang, the protocol’s head of growth, maintains a close advisory relationship with WorldClaw.
The national security dimension #
Both WLFI and WorldClaw have pushed back on this framing. Their position is that other major US technology firms routinely work with Chinese AI developers and that this kind of cross-border model aggregation is standard industry practice.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our