The United States is no longer treated as a sovereign power but as a distressed asset. Simon Dixon on Hard Talk argues that global capital, steered by AI, now sees the US as a debt-laden entity to be liquidated. Firms like BlackRock use algorithmic models to shift capital from struggling US regions to higher-return zones like West Asia. The result: the state keeps the debt, while private interests take the profits.
Dixon claims Beijing holds a financial and technological kill switch. China’s holdings of US debt - though reduced to $650 billion - could still destabilize markets if dumped. More critically, US weapons systems rely on Chinese rare earths and components. A supply cutoff would paralyze defense production. This isn’t speculation; it’s leverage in plain sight.
"China doesn't need to fire a shot. They can just stop selling the magnets and chips that keep the American machine running."
- Simon Dixon, Hard Talk
Meanwhile, US policy is accelerating dependency on Chinese tech. Howard Lutnick’s AI access regime restricts frontier models like Mythos 5 and GPT-5.6 to a narrow government-approved circle. This bottleneck is pushing companies offshore. Coinbase now defaults to Chinese open-weight models like GLM 5.2 and Kimmy 2.7 to cut AI costs in half.
Open Router’s June report confirms a shift: DeepSeek v4, Qwen 2.7, and GLM 5.2 are now standard in production agentic workflows. The Wall Street Journal found GLM 5.2 matches Mythos in bug detection. Analysts warn the US is creating a self-inflicted intelligence gap - hoarding top-tier AI while pushing allies toward Chinese stacks.
"The world on the other side of these bans will be fundamentally different."
- Nathaniel Whittemore, The AI Daily Brief
The pattern is clear. US restrictions weaken domestic competitiveness. China fills the void with open, capable models. The state may believe it’s protecting security - but it’s actually ceding ground. The future isn’t just multipolar. It’s being built on Chinese code.

