"Techno-Federalism: How Regulatory Fragmentation Shapes the U.S.-China AI Race" is an academic article by Jason Jia-Xi Wu, published in 2025 in the Harvard National Security Journal, Vol. 17, No. 1 (cited as 17 Harv. Nat'l Sec. J. 1 (2025); posted on the journal's site in December 2025) (Source: journals.law.harvard.edu). The article argues that the United States and China are converging toward a fragmentary approach to AI governance in which the tech industry acts as a co-regulator alongside central and local governments, and that this convergence challenges the framing of the AI race as a "battle of values" between liberal democracy and techno-autocracy. The piece runs roughly 80 pages with extensive footnotes. At the time of writing Wu was a judicial law clerk at a U.S. District Court and held a J.D. from Harvard Law School.
Summary of argument
Wu contends that the US-China AI race is not merely a contest of values but a movement by both countries toward a shared middle ground from opposite ideological starting points. He labels the resulting arrangement "techno-federalism," a term combining "technocracy" with "federalism" to describe the tripartite interplay of central government, local government, and market power in AI governance. In this framing the tech industry functions as a co-regulator: not simply a regulated party but an actor that sets de facto norms alongside the state.
The convergence thesis holds that the United States is departing from neoliberal free-market politics and moving toward national security objectives, while China is moving from state corporatism with centralized control toward decentralized innovation that blends national security with local experimentation. In the United States, states lead AI regulation while the federal government takes a backseat, and tech firms act as co-regulators through self-governance, lobbying, and regulatory arbitrage. In China, local governments implement and experiment with national AI policies, and tech firms act as co-regulators that implement national strategic objectives with local discretion. The two systems and the two directions of movement are summarized below.
| Dimension | United States | China |
|---|---|---|
| Starting point | Neoliberal free-market politics | State corporatism with centralized control |
| Direction of movement | Toward national security objectives, departing from free-market principles | Toward decentralized innovation, blending national security with local experimentation |
| Role of states/localities | States lead AI regulation; federal government takes backseat | Local governments implement and experiment with national AI policies |
| Role of industry | Tech firms as co-regulators via self-governance, lobbying, regulatory arbitrage | Tech firms as co-regulators implementing national strategic objectives with local discretion |
Three domains of governance
Wu analyzes techno-federalism across three governance dimensions. In software governance, California leads in regulating frontier model training and data aggregation; in China the Cyberspace Administration sets national rules but local enforcement varies; and industry sets de facto norms through model releases, safety practices, and open-source initiatives. In hardware governance, Arizona leads US semiconductor development through manufacturing subsidies tied to CHIPS Act implementation, China's semiconductor policies mix central direction with local incentive competition, and Export Controls (AI) are a key US tool whose implementation industry shapes through lobbying and supply chain decisions. In infrastructure governance, New York and Delaware dominate AI infrastructure financing through corporate and commercial law, China's datacenter and cloud infrastructure policies balance local protectionism with national coordination, and energy, datacenter siting, and networking are governed by a patchwork of local regulations.
Relation to traditional federalism
Wu argues that techno-federalism differs from traditional federalism in three respects. It is not the product of constitutional design but emerged organically in response to AI's evolving landscape, creating blurred regulatory boundaries. It is marked by legal uncertainty, with no clear allocation of which entities are responsible for which aspects of AI governance, in contrast to traditional federalism's cleaner divisions. And it is shaped by market norms, with tech firms operating under a patchwork of state and local laws and effectively setting standards through self-governance.
Wu compares the current competition to the Cold War "triple helix" of government, academia, and industry but argues that today's version is fundamentally different. In the Cold War the arrangement was government-led, with academia and industry operating as extensions of the military-industrial complex on the DARPA model. Today, by Wu's account, industry leads AI development and often dictates norms without federal oversight, the boundaries between all three actors are fluid, and regulatory fragmentation enables states and industry to act as co-regulators.
Policy recommendations
For state governments, Wu recommends that states coordinate on AI governance to reduce regulatory fragmentation while preserving the "laboratory" benefits of state-level experimentation, leading in software, hardware, and infrastructure governance within their respective comparative advantages. For the federal government, he recommends developing a coherent federal AI policy that aligns national security objectives with commercial incentives, addressing the patchwork of inconsistent state AI regulations without resorting to full preemption, and harnessing industry self-governance while maintaining checks on private-sector power. The article's Part V explores the policy implications of techno-federalism, examining its impact on the future trajectory of the U.S.-China AI race, and proposes legal reforms along these lines (Source: journals.law.harvard.edu).
Relation to other work
Wu's framework bears on several adjacent debates. It complicates the Eight Worlds Framework by suggesting that Sullivan and Feldman's axis of "China racing vs. not racing" is too simple, since on Wu's account China's AI strategy is itself fragmented between central direction and local experimentation. It nuances AI and Authoritarianism by arguing that the CCP's AI governance is not monolithically centralized, with local governments and private firms holding significant discretion, a more complex picture than Amodei's framing suggests. And it extends AI Diffusion by treating diffusion as not only international, between US and Chinese systems, but also intra-national, between federal and state and between central and local levels.
The regulatory fragmentation Wu describes is the context in which the AI LEAD Act operates, as a federal liability law attempting to create coherence amid a patchwork of state regulations. State-led measures such as SB 53 and the RAISE Act are examples of the state-level AI regulation his framework predicts and explains. The algorithmic pricing antitrust debate illustrates how different legal frameworks — state versus federal, antitrust versus product liability — can produce conflicting regulatory signals.
Relationships
- supports: Techno-Federalism
- related: Eight Worlds Framework, AI and Authoritarianism, AI Diffusion, AI LEAD Act (S. 2937), California SB 53 — Transparency in Frontier AI Act, New York RAISE Act (S. 8828), Algorithmic Pricing and Antitrust, Export Controls (AI)
Provenance
The article appears in Volume 17, Issue 1 of the Harvard National Security Journal, published on the journal's site in December 2025; the full text is available from the journal's landing page (Source: journals.law.harvard.edu) and as the journal's typeset PDF (Source: journals.law.harvard.edu). This summary derives from the article text preserved in Raw Sources/Techno-Federalism - How Regulatory Fragmentation Shapes the US-China AI Race.md.