August 16, 2025: Position Before Force
Leverage is often described as a matter of strength, but it is just as much a matter of position. A tool’s force depends on what it rests against — a structure that holds while pressure is applied. In political economy, those fixed points are consistency, time, and alignment. Without them, even the most sophisticated instruments can move without shifting anything of consequence.
This is the difficulty of using export controls as a strategic tool. They are designed to alter another state’s capabilities over years, not weeks. Their effect depends on predictability — the belief that restrictions will endure long enough to influence investment decisions, research priorities, and market behaviour. When they change quickly, they lose this quality, becoming signals of intent rather than mechanisms of constraint.
In August, the United States reversed a three-month-old ban on the export of Nvidia’s downgraded H20 artificial intelligence chips to China. These chips were engineered to comply with earlier export restrictions by limiting their performance while still meeting commercial demand. The decision to restore sales came with a fifteen percent levy on revenue, negotiated directly between the president and Nvidia’s chief executive, Jensen Huang. The arrangement could yield $3 billion on $20 billion in projected sales — a sum small in fiscal terms but immediate and concrete enough to headline the announcement.
The public framing emphasised the deal-making, not the strategic function of the policy. The stated goal of restricting advanced AI chips to China has been to slow its ability to develop systems with military and security applications, preserving a technological lead for the United States. Whether the H20 decision advances or undermines that aim depends on its place within a sustained framework. As a one-off bargain, it trades a source of potential leverage for a modest financial gain.
China’s own use of export controls offers a longer view. For more than a decade, it has restricted the export of rare earths and other critical inputs, spanning hundreds of product categories and tracking end-users. These measures are integrated into broader industrial and political objectives, designed to be sustained and to influence behaviour over time. Whatever their limitations, they are coherent. The United States’ handling of the H20s, by contrast, appears episodic, responsive to the needs of a single negotiation rather than anchored in a continuous strategy.
This difference reflects more than contrasting political systems. In China, the integration of major firms into the state’s industrial framework ensures alignment between corporate action and national objectives. Party oversight cells operate inside companies, and high-profile executives can be removed from public life when they challenge political priorities, as with Alibaba’s Jack Ma in 2020. In the United States, corporate cooperation is negotiated on a case-by-case basis. Political calls for executive resignations, such as those directed at Intel’s Lip-Bu Tan and Goldman Sachs’s David Solomon in the week before the H20 announcement, carry no binding authority. Alignment, when it happens, is temporary.
Export controls applied in such a setting are inherently fragile. They rely on the willingness of private firms to accept limits that may conflict with their market incentives, and on a political environment willing to maintain those limits over time. Reversals weaken credibility. In resource markets, lifting an embargo prompts adaptation — alternative suppliers, new processes — making future restrictions less effective. In technology, the process is slower, but the outcome is similar. Every period of access reduces the distance the restriction was meant to preserve.
The $3 billion payment from the H20 arrangement is measurable, immediate, and easy to claim as a success. The value of sustained leverage over technological development is harder to quantify and less visible in the short term. This difference in timescale and visibility often tilts policy toward the short-term gain, especially when decisions are not bound into a larger, enduring plan.
What this reveals is not just the limits of one policy decision, but a recurring feature of U.S. economic statecraft: the ability to act without the conditions needed to make action effective over time. Tools exist — export controls, subsidies, tariffs — but without the fixed points of consistency and alignment, their movement does not always shift the structure beneath. The open question is whether those points can be built in a system that depends on the voluntary cooperation of private actors for much of its strategic capacity.