Limitations of Export Controls, A Thought Exercise

Over the fall semester of 2025, I researched the importance of South Korean technology firms as an underappreciated node of the semiconductor supply chain. The findings and recommendations for export controls were posted in a Substack article titled South Korea and the HBM Bottleneck, arguing for export controls that were more predictable and sensitive to structural constraints that South Korean firms faced.

The preceding analysis treats export controls and supply-chain bottlenecks as meaningful constraints in U.S.–China competition on frontier AI development. While this framing reflects dominant policy discourse, I remain uncertain about several of its underlying assumptions.

My training is primarily in political sociology rather than trade or security studies; thus, these notes reflect an exercise to think across disciplinary boundaries more so than to offer wholesale policy prescriptions.

Rethinking the Role of Export Controls

One line of skepticism concerns the effectiveness of export controls: whether they meaningfully constrain long-run access to compute. Washington, across three administrations, has imposed controls on the export of high-end semiconductors, chipmaking equipment, and related computer systems to China and a list of Chinese entities. In the short term, these measures did disrupt Chinese compute capacity and introduced real frictions for firms attempting to scale.

However, despite these constraints, evidence suggests that Chinese firms retain multiple pathways for accessing compute, including smuggling, cloud intermediaries, third-country channels, and stockpiling during periodic regulatory relaxation. Chips are small, high-value, and easy to conceal, making undetected cross-border transport comparatively facile through third-country intermediaries and offshore arrangements. Even where enforcement is strict, firms retain access to compute via cloud intermediaries operating outside U.S. jurisdiction, including in neighboring countries such as Japan. These dynamics raise questions about whether enforcement and monitoring costs scale proportionally with long-term effectiveness.

In early October of 2025, I expected that Beijing would forgo renewed access to U.S. accelerators in favor of accelerating domestic substitution. The reopening of Nvidia sales—and China’s willingness to purchase—proved that expectation wrong. At the same time, domestic chip production capacity continues to scale, even as China reportedly places an order for two million H200 chips that Nvidia is eager to fulfill. This episode underscores how access constraints function less as binding limits than as temporary frictions layered onto an industrial system that continues to expand along multiple pathways.

Even if export controls succeed in delaying capability acquisition, they remain a blunt and adversarial instrument when treated as a substitute for industrial or diplomatic strategy. Rather than inducing compliance, they often incentivize adaptation. In this sense, export controls resemble interdiction policies more than structural solutions. Sociological scholarship on drug enforcement has long described the “balloon effect” and the “hydra effect,” wherein suppressing flows in one channel results in their emergence elsewhere as long as underlying demand remains intact. While the analogy is imperfect, a similar dynamic appears in technology controls: demand for compute remains high, and restrictions tend to redirect supply rather than eliminate it.

These concerns are compounded by questions of political sustainability under the current U.S. political environment, which is characterized by executive volatility and conflict with the federal bureaucracy. For firms and allied governments whose compliance is essential to enforcement, export controls introduce persistent uncertainty. This uncertainty incentivizes hedging behavior: stockpiling during regulatory relaxation, accelerated investment in alternative supply chains, and diversification away from U.S.-centered technological dependencies.

When applied to a large industrial state such as China, volatile enforcement regimes further encourage prioritization of domestic substitution and long-term capacity building. Signals that access to critical technologies cannot be relied upon—even during periods of nominal détente—accelerate endogenous adaptation. Over time, this dynamic may reinforce precisely the self-sufficiency and decoupling trajectories that export controls are intended to slow.

The HBM bottleneck illustrates a related displacement effect common to sanctions and export-control regimes. Constraints imposed on the target state are partially externalized onto allied firms whose participation is necessary for enforcement but costly in practice. South Korean firms, occupying a critical node in the semiconductor supply chain, bear significant commercial and planning risks under shifting regulations, even as the long-term strategic payoff of these constraints remains uncertain. Their experience highlights how export controls redistribute costs across the alliance network rather than containing them at the target alone.

From a political-economy perspective, technology controls can be understood as attempts to stabilize existing global technological hierarchies through administrative restriction rather than through transformation of underlying productive relations. Such strategies appear structurally fragile when applied to industrial states with sufficient capital, labor, and state capacity to reorganize production domestically. In this sense, export controls may function less as durable instruments of maintaining incumbent power than as transitional mechanisms that accelerate multipolar reconfiguration.

At the same time, skepticism toward export controls should not be overstated. Delays, frictions, and uncertainty can matter in fast-moving technological domains, particularly where scale effects and learning curves apply. The question, rather, is whether export controls can bear the strategic weight often placed upon them—or whether longer-term industrial coordination and diplomatic strategy are required to complement them.

What This Means for the HBM Argument

These notes are not an argument against export controls per se, but a record of growing uncertainty about their long-term efficacy and broad consequences. They suggest that such mechanisms may function best as temporary stabilizers rather than as a durable foundation for long-term competition with an industrial power of China’s scale. Concentrated production and technical specialization can create short-term leverage, particularly where allied coordination is credible. At the same time, the uncertainty borne by allied firms underscores the risks of relying on chokepoints as a primary strategic instrument over the long run. I offer these points in the spirit of clarification rather than conclusion.