“30 Days Isn’t Strategy”: Quant Model Calls for 500 t Dysprosium + 150 t Terbium to Protect U.S. Defense Continuity

Highlights

• Pokorny's 2026 study reveals a 22-fold gap: the National Defense Stockpile's ~30-day HREE coverage falls critically short of the modeled optimal 608-730 days, with China's supply chain dominance creating acute national security vulnerability for defense magnets.
• Two-stage stochastic optimization across 10,000 Monte Carlo simulations recommends stockpiling 500 metric tons of dysprosium and 150 metric tons of terbium at $370M cost, projecting 96% defense continuity and 75.5% zero-shortage probability over 10 years.
• The quantitative framework balances acquisition, storage, and shortage penalty costs under four disruption scenarios, though implementation questions around material form, rotation protocols, and thin-market acquisition timing remain unresolved.

---

The United States faces a national security weak point hiding in plain sight: heavy rare earth elements (HREEs)—especially dysprosium (Dy) and terbium (Tb)—remain overwhelmingly foreign-supplied and are difficult to replace in high-performance magnets used across defense platforms. In “Strategic Stockpile Optimization for Heavy Rare Earth Elements” (2026), Laszlo Pokorny, ICL Institute, Department of Strategic Resources and National Security, New Jersey, USA, presents a quantitative framework arguing that the National Defense Stockpile’s estimated ~30 days of HREE coverage is far below what prudent planning demands. Using two-stage stochastic optimization integrated with Monte Carlo simulation (n=10,000) and four disruption scenarios, Pokorny estimates “optimal” stockpile targets of 500 metric tons of dysprosium and 150 metric tons of terbium, requiring roughly $370 million in upfront acquisition and delivering 608–730 days of supply coverage over a 10-year planning horizon—alongside a modeled 96% defense continuity rate and 75.5% probability of zero shortage events.

Study Approach

Pokorny builds a decision model that tries to answer: How much Dy and Tb should the U.S. hold to minimize total expected costs while avoiding crippling shortages? The model weighs three cost buckets—buying the material, storing it, and the penalty of being short during a disruption—under multiple disruption severities. Inputs draw from USGS, IEA, and UN Comtrade over a 2010–2024 historical window.

Key findings and why they matter

• A “22-fold gap”: the study asserts today’s coverage (~30 days) is far below modeled needs (roughly 1.7–2.0 years).
• China-linked concentration risk: the study frames China’s dominance in the value chain as the core vulnerability—especially for HREE separation/processing and magnet supply.
• Stockpiling as deterrence and insurance: the model’s outputs quantify the cost–security trade-off and argue forimmediate recapitalization.

 Limitations and Debate Points

This is a model-based, scenario-driven study (not operational DoD planning guidance and not presented here as peer-reviewed validation). Results depend heavily on assumptions about shortage costs, disruption probabilities, and demand. The framework focuses on Dy and Tb only, and does not fully resolve critical implementation questions (e.g., what form to stockpile—oxide/metal/alloy, quality specs, rotation, storage constraints, and acquisition timing impacts on thin markets).

Author profile

Dr.Laszlo Pokorny is an educational technology consultant with 20+ years inK-12 and higher education. He holds an EdD in Educational Technology Leadership, an MS in Pharmaceutical Economics, an MTeach in Special Education, and a BS in Plant Molecular Biology, with NJ teaching certifications in biology, physics, and special education. He founded the ICL Institute (2022) for applied interdisciplinary research and the Post-Transplant Research Institute (PTRI) (2024) following his heart transplant.

Citation: Pokorny, L. (2026). Strategic Stockpile Optimization for Heavy Rare Earth Elements: A Quantitative Framework for National Security Planning (opens in a new tab). ICL Institute. https://doi.org/10.5281/zenodo.18522925 (opens in a new tab)