Counterintuitively, U.S. and global auto firms need to collaborate with Chinese suppliers to reduce strategic dependency. The model involves onshoring Chinese hardware and manufacturing expertise while maintaining national control over sensitive AI software and networks, creating a strategic "co-opetition."

A "chips for rare earths" equilibrium underpins the U.S.-China relationship. Any concession by China on rare earth controls would likely require reciprocal U.S. flexibility on advanced semiconductor exports. This strategic deadlock disincentivizes escalation from either side on the technology front.

Current US policy is reactive, fixing compromised supply chains like semiconductors. A proactive 'offensive' strategy would identify nascent, critical industries (e.g., humanoid robotics) and build the entire supply chain domestically from the start, securing a long-term economic and national security advantage.

The current trade friction is part of a larger, long-term bipartisan U.S. strategy of "competitive confrontation." This involves not just tariffs but also significant domestic investment, like the CHIPS Act, to build resilient supply chains and reduce reliance on China for critical industries, a trend expected to persist across administrations.

Selling semiconductor equipment allows China to create hundreds of billions in downstream value. In contrast, selling API access to US models is a higher-margin strategy that keeps core value creation within the American ecosystem, extracting more revenue per unit of capability provided.

The shift towards a less aggressive stance is not weakness, but a strategic pause. Both the U.S. and China need time to build domestic strength, creating a temporary 'modus vivendi' the U.S. can use to improve its long-term competitive position.

A zero-tolerance policy on selling advanced AI chips to China might be strategically shortsighted. Allowing some sales could build a degree of dependence within China's ecosystem. This dependence then becomes a powerful point of leverage that the U.S. could exploit in a future crisis, a weapon it wouldn't have if China were forced into total self-sufficiency from the start.

The most effective way to prevent conflict between the US and China is to create mutual, bidirectional economic dependency. This involves significant US exports (planes, cars, chips) into China's consumer market, balancing the historical one-way flow of cheap goods and moving beyond political posturing.

A complete ban on selling chips to China is counterproductive. The ideal policy allows NVIDIA to sell chips that are one or two generations behind state-of-the-art. This strategy keeps Chinese firms dependent on the NVIDIA ecosystem, funds U.S. R&D with sales revenue, and hinders domestic competitors like Huawei from flourishing.