Highlights
- A Yale study reveals the Trump administration's 2025 AI Action Plan aims to accelerate innovation and maintain global AI leadership, but faces significant supply chain challenges.
- The global AI ecosystem remains tightly interconnected, with China dominating rare earth production and the U.S. dependent on international chip fabrication.
- Decoupling AI technologies between the U.S. and Chinese blocs is structurally unattainable due to deep technological and mineral supply interdependencies.
In a recent paper, Emmie Hine (opens in a new tab) (Yale Digital Ethics Center, University of Bologna, KU Leuven) and Luciano Floridi (opens in a new tab) (Yale University, University of Bologna) dissect the Trump administration’s July 2025 AI Action Plan and its supporting executive orders. The aim according to the researchers? Accelerateinnovation, build American AI infrastructure, and lead global AI diplomacy. While framed as a bold new vision, the authors argue the plan largely continues prior U.S. efforts to maintain competitiveness against China, now layered with distinctly Republican cultural and political priorities.
Most crucial for investors and supply-chain watchers: the study underscores that deep interdependencies in rare earths and semiconductors render full AI “decoupling” between U.S. and Chinese blocs nearly impossible. China still mines ~70% and processes ~90% of global rare earths, while U.S. chip giant Nvidia commands a 92% GPU market share but relies on Taiwan’s TSMC for fabrication. This dual chokehold keeps the world’s AI ecosystem bound together despite combative rhetoric.
Implications for Rare Earths and AI
For rare earth investors, this report is a reminder that Washington’s AI strategy is inseparable from critical minerals. AI data centers, chip fabs, and defense-linked computing infrastructure all depend on steady flows of neodymium, dysprosium, terbium, and other rare earths. While the Action Plan outlines big ambitions for U.S. chip sovereignty, rare earth policy is notably downplayed—suggesting it may be relegated to quieter, less visible industrial initiatives.
The takeaway: markets should expect heightened government support for U.S. rare earth projects, but with limited ability to break China’s dominance in the medium term—just as Rare Earth Exchanges (REEx) has forecasted. Investors should watch for strategic stockpiling, subsidy expansion, and cross-sector deals (as seen in the July 2025 U.S.–China rare earth–Nvidia trade linkage).
The Limits of the Plan
Hine and Floridi highlight several structural and political constraints:
| Challenges | Summary |
|---|---|
| Supply chain knots | Rare earths and chips remain too globally interdependent for clean bloc separation |
| Legal challenges | New ideological neutrality mandates for AI may face constitutional tests |
| Alliance friction | Past U.S. tech diplomacy (e.g., on Huawei/5G) alienated allies, raising doubts about sustaining an “AI alliance.” |
| Industrial lab | U.S. onshoring of chipmaking won’t mature until at least 2026–2027, keeping reliance on Taiwan intact. |
Conclusion
The study concludes that while the AI Action Plan marks the clearest articulation yet of the Trump administration’s ambitions, rare earth supply chains remain the Achilles’ heel of U.S. technological independence. Decoupling may be politically popular but structurally unattainable. For rare earth investors, this underscores both the opportunity—expanded U.S. support for domestic mining and processing—and the enduring reality of China’s leverage.
Citation: Hine E, Floridi L. Winning the AI Race? The US AI Action Plan in Context. Yale Digital Ethics Center (opens in a new tab); University of Bologna; KU Leuven. September 2025.
©!-- /wp:paragraph -->
0 Comments