We have arrived at a moment in technological affairs that bears striking resemblance to the great constitutional debates that preceded the framing of the Constitution itself—a moment when competing sovereignties, each claiming legitimacy to direct the development of critical infrastructure, threaten to produce chaos rather than coordination. The global semiconductor industry, once governed by market forces and competitive advantage, has now become the primary theater of what scholars call "great-power competition," wherein governments worldwide deploy unprecedented fiscal, regulatory, and geopolitical instruments to secure domestic AI and chip supply chains. For observers of institutional design, this transformation presents a crucial case study in how concentrated technological power, when left ungoverned by clear constitutional boundaries, invites the very abuses of authority we sought to prevent.
The paradox before us is profound: governments everywhere recognize that semiconductor and artificial intelligence capacity is synonymous with national sovereignty itself. Yet their remedies—vast subsidies, export prohibitions, equity stakes in private enterprises, and increasingly baroque compliance regimes—risk fragmenting the global technology commons into competing fiefdoms, each governed by opacity and discretion rather than law. As we examine the major governmental initiatives now reshaping this landscape, we do well to ask a foundational question: What are the least dangerous institutional arrangements by which democracies can secure technological self-sufficiency without surrendering the rule of law itself?
The Scale and Scope of Government Intervention: A Historical Recalibration
The genius of modern market capitalism lies in its assumption that competition and price discovery serve the national interest better than targeted state direction. Yet the semiconductor industry has shattered that assumption. Consider the sheer magnitude of government commitment now deployed.
Japan has committed approximately $2.3 trillion over 14 years toward semiconductor production and vertical AI applications 3,5—a figure corroborated across multiple independent sources and therefore carrying high confidence. The United States, through the CHIPS and Science Act, has authorized $52 billion in grants and tax credits 13, and this federal commitment has already catalyzed over $450 billion in private semiconductor investment 10. India has approved 12 semiconductor projects worth approximately ₹1.64 lakh crore and launched its Semiconductor Mission 2.0 to encompass the full value chain 32. The European Union is advancing Chips Act 2.0 to stimulate demand and bridge the persistent gap between research and market-ready production 9. China's "Big Fund" targets 50 percent domestic semiconductor production across all sectors 8. Collectively, these initiatives establish trillion-dollar investment bills as the new baseline for national semiconductor self-sufficiency 3.
What we are witnessing is a deliberate departure from the post-1945 framework of comparative advantage and integrated global supply chains. Instead, each major power has elected to replicate, within its own borders, capabilities that were once distributed across geography and optimized by market forces. This is not, by itself, irrational—national survival does sometimes supersede economic efficiency. But it represents a fundamental restructuring of the technological commons and a corresponding multiplication of the sites at which government authority is exercised over private enterprise.
The American Gatekeeper: Export Controls and the Concentration of Authority
Here we confront a question analogous to those the Framers debated in 1787: When power concentrates in a single authority—in this case, the U.S. government's control over processor exports—what mechanisms prevent its abuse? The answer, as yet, remains unclear.
Since October 2022, the United States has prohibited exports to China of processors at or above the performance level of the NVIDIA A100 14. But the restrictions have evolved in scope and ambition. The Trump administration is now drafting rules that would require U.S. government approval for exports of NVIDIA and AMD AI chips to every country, including close allies such as Britain, Germany, and Japan 29—a measure corroborated by multiple independent sources. Simultaneously, a January 2026 Presidential Proclamation imposed a 25 percent duty on advanced computing chips under Section 232 Phase 2 21,22, though products destined for data-center use were explicitly exempted 21.
The proposed MATCH bill would expand export restrictions to include deep ultraviolet immersion equipment—the very machinery upon which semiconductor fabrication depends 15. In substance, the Bureau of Industry and Security is positioning the U.S. government as a gatekeeper for nearly every high-end processor sold by an American firm 18. Historical approval timelines for AI chip export licenses have averaged approximately one year 11, a duration that creates significant revenue uncertainty for any company dependent on international sales.
This is the modern form of mercantilism—not tariffs on finished goods, but an administrative licensing regime that vests discretionary authority in executive agencies. As Madison observed in Federalist No. 10, a republic must be designed to control the effects of faction; yet when a single agency holds the power to grant or deny market access across the globe, the potential for corruption and rent-seeking becomes substantial. We have created, perhaps inadvertently, a system in which technological competitiveness is no longer primarily a matter of innovation or manufacturing excellence, but of proximity to state power.
Enforcement, Evasion, and the Bifurcation of Global Markets
The concentration of such authority in the executive branch necessarily invites evasion and creates incentives for circumvention. The evidence is now materializing.
Taiwan conducted its first public crackdown on AI chip diversion in response to years of U.S. pressure 25, and Taipei is considering criminalizing unauthorized AI chip exports 25. Malaysia has emerged as a suspected smuggling intermediary, with reports alleging that U.S.-manufactured AI chips were routed to China via Malaysian entities 20, prompting permit requirements under the Malaysian Strategic Trade Act 20. The Center for a New American Security estimates that between 10,000 and several hundred thousand AI chips were smuggled into China during 2024 alone 14. These enforcement actions directly threaten the ability of American firms to monetize demand in restricted markets and could accelerate the bifurcation of the global AI chip market into a Western sphere and a Chinese sphere, each governed by incompatible standards and supply chains.
This pattern is instructive for students of constitutional design. When a rule is perceived as fundamentally unjust—as an arbitrary restraint on commerce that benefits one nation's competitors—it generates exactly the sort of resistance that undermines both the rule and the authority that established it. The genius of the Constitution lies partly in its acceptance that you cannot govern a large republic through absolute prohibition; you must instead build systems of consent and distribute power such that no single faction can impose its will upon the whole.
Sovereign Equity and the Distortion of Corporate Governance
A second and perhaps more insidious mechanism of government control has emerged: the use of public equity stakes to reshape corporate governance within nominally private firms.
The CHIPS Act couples subsidies with equity stakes in private manufacturers 30. The U.S. government has taken an equity position in Intel, backed by $8.9 billion in CHIPS Act proceeds and Secure Enclave funding 27,28. The Vice President has expressed support for sovereign-wealth-style equity stakes in major AI companies 7, and proposed acquisitions of equity in AI labs may require congressional authorization 17. For enterprises dependent on semiconductor supply, this creates a novel uncertainty: if the U.S. government becomes a meaningful shareholder in a domestic fab or chip designer, will procurement decisions be made on commercial merit or on grounds of national interest? Will technology-sharing arrangements be mandated? Will margins be compressed through preferential pricing imposed by a shareholder-state?
This is a problem less of democracy than of institutional clarity. The Framers understood that when you combine the power to regulate an industry with a financial interest in that industry's success, you create irreconcilable conflicts of duty. A shareholder cannot serve two masters—the return on investment and the public interest—especially when those two goods diverge. Yet we are now accumulating precisely such conflicts in the semiconductor sphere. The CHIPS Act's emphasis on front-end wafer fabrication over advanced packaging may also create bottlenecks in the very packaging technologies that advanced chiplets depend upon 13, inadvertently damaging the competitive position of American firms that rely on sophisticated packaging to maintain their performance lead 33.
The Regulatory Patchwork and the Problem of Jurisdiction
Layered atop these federal mechanisms sits an increasingly fractured regulatory regime at the state level, a problem that would have been immediately recognizable to delegates to the 1787 Convention.
Approximately 100 AI-related legislative measures have been introduced or enacted across 38 U.S. states 1, with states like Illinois, Texas, Colorado, and California advancing governance laws that exceed the federal baseline 2. This creates a patchwork incompatible with the unified nation-state that both commerce and security require. The Great American AI Act, introduced in the House, seeks to preempt state regulation of AI model development while permitting states to regulate AI use 4,26—a provision corroborated across multiple sources. The White House has made preemption a federal legislative priority 6. Yet the U.S. currently lacks comprehensive federal AI regulation 31, and the executive branch has shown a preference for directives over formal legislation 16.
This is not merely a problem of administrative efficiency. It is a fundamental question of constitutional authority: In what body—federal or state—should the power to regulate AI reside? And if the federal government preempts state authority over AI development while states retain authority over AI use, have we created a coherent rule or a recipe for conflict? The Constitution's Commerce Clause emerged precisely because the states' attempts to regulate trade locally produced chaos. A well-constructed framework must balance the need for uniform rules over interstate commerce with the principle of subsidiarity—the notion that decisions should be made at the smallest competent level of government. We are not yet there.
China's Response: Domestic Self-Sufficiency and Technological Circumvention
No account of global semiconductor governance would be complete without acknowledging the feedback loop that American export controls have created. China is not passively accepting its exclusion from advanced Western chips; it is investing billions in domestic self-sufficiency and pursuing technological innovations designed to circumvent Western constraints 19,24.
Chinese firms are advancing 3D stacking innovations to compensate for performance disadvantages relative to American chips 12. Estimates suggest that China could manufacture 200,000 advanced AI chips in 2025—a small fraction of global output, but a directional trend pointing toward competitive independence 23. Over a decade or two, the convergence of sustained Chinese investment, American export restrictions, and technological learning-by-doing may produce a scenario in which China possesses a fully functional domestic AI infrastructure, independent of American components. The unintended consequence of attempting to exclude a nation from markets through export control is often to catalyze the very self-sufficiency you sought to prevent.
The Question of Financing Viability
Finally, we must confront a question that governments are reluctant to ask: Is the cost of achieving domestic semiconductor self-sufficiency economically sustainable?
One claim warns that CHIPS Act funding in both the U.S. and the EU may not be sufficient to offset long-term operational cost differentials, creating uncertainty about the commercial viability of Western semiconductor investments 8. If subsidized fabs prove uneconomic without perpetual government support, the downstream demand for advanced chips in those facilities could be at risk. This raises a constitutional dilemma: Can a government that has committed to subsidizing an industry indefinitely remain, in any meaningful sense, a limited government constrained by enumerated powers?
Implications for Technology, Trade, and National Governance
The cluster of policies now arrayed across the semiconductor industry—export controls, equity stakes, subsidies, state-level regulations, and enforcement mechanisms—presents a profound challenge to the liberal internationalist order that has governed trade and technology since 1945. That order rested on the assumption that markets would allocate resources efficiently and that competition would spur innovation. The emerging alternative assumes that the state must direct technological development and that national sovereignty supersedes comparative advantage.
This is not, in itself, illegitimate. Nations do have the right to secure their defensive capabilities. But the mechanisms now being employed—opaque licensing, concentrated executive authority, private equity held by the state, and fragmented regulatory regimes—threaten to erode the very rule of law that distinguishes liberal democracies from autocracies. The great danger here is the accumulation of unchecked authority in semiconductor gatekeeping, with minimal parliamentary or judicial oversight.
A well-constructed alternative framework might rest on several principles. First, export controls should be narrowly tailored to genuine security threats, defined ex ante in legislation rather than through executive discretion. Second, if the state takes equity stakes in private firms, that stake should be passive and constrained by law; no sovereign shareholder should exercise control over hiring, procurement, or technology decisions. Third, the regulatory patchwork at the state level should be preempted by a clear federal baseline, but only after a deliberate legislative process—not through executive fiat. Fourth, alliances with democracies should feature transparency regarding technological restrictions and reciprocal agreements on supply-chain cooperation. And fifth, subsidies for semiconductor fabrication should be time-limited, subject to sunset provisions, and evaluated regularly against commercial benchmarks to prevent permanent dependency.
The semiconductor industry is not a simple policy question. It implicates security, innovation, employment, and international order. But it is manageable if—and only if—we approach it with the same rigor that the Framers brought to the structure of government itself: with clear enumeration of powers, mutual checks and balances, transparency, and a stubborn insistence that concentrated authority must justify itself in law, not in the exigencies of the moment.
