When can independent capability deliver?
Test when available material becomes an accepted NdFeB component. Vary downstream availability and qualification against a specified delivery deadline.
Start here / 从这里开始 · Four chapters
Follow the magnet
through four chapters.
Begin with demand and applications. Follow the mining origins and materials, enter the factory, then explore why future supply matters.
Start 01 · Demand / Applications→从第一章开始 · 需求流量与应用- 01
Demand / Applications
需求流量 / 应用Read sourced magnet demand and the requirements of the machines that use it.
Explore demand ↗ - 02
Supply
供应Follow mining origins, compare grades and formulations, and inspect material prices.
Follow the supply ↗ - 03
Factory Tour
工厂之旅Watch materials change through production steps and examine equipment readiness.
Enter the factory ↗ - 04
Future Importance
未来的重要性Explore research questions, investment, supply security and scenario laboratories.
Explore the future ↗
From earth to motion
Where does
resilience break?
More mines do not automatically mean more usable magnets. Explore the passages between a mineral resource and a machine that can finally move.
Enter the supply chainNeodymium & dysprosium · a preliminary physical supply model
THE NINE-PASSAGE SYSTEM
FINITE STOCKSREAL PASSAGE OF TIME
DESIGNED TO QUESTION ASSUMPTIONSSCROLL TO INVESTIGATE ↓MODEL v0.1 / OCT 2026
01 / The research lens
A resource is not
yet a capability.
Three questions turn a concentration statistic into a testable supply-chain problem.
↗
When does the preferred intervention change?
Explore a matched abstract budget. Inventory is pre-positioned; capacity starts after a decision and qualification lag. Their different availability dates affect the comparison.
When does a buffer become a bottleneck?
Vary shock duration, stock cover and qualification time to see when a temporary interruption becomes delayed deployment.
02 / The supply-chain board
Nine passages.
One connected system.
Choose a passage to inspect what happens there. Nine explanatory operations map to three aggregate production groups. The experiment divides capacity between China and the rest of the world.
Selected passage Disrupted model groupNine explanatory operations ≠ nine separately calibrated production models.
03 / The live experiment
Change the conditions.
Follow the consequences.
LOCAL SIMULATIONConditional results, not a forecast
START WITH A QUESTION
THE SELECTED SCENARIO
Refining interruption60NdNEODYMIUM
—contained tonnes / yearWind + PM vehicles
66DyDYSPROSIUM
—contained tonnes / yearWind + PM vehicles
DELIVERED BY HORIZON—of cumulative portfolio demand
PEAK DEPLOYMENT BACKLOG—months of the fixed portfolio
DELAY BURDEN—portfolio-months × months delayed
WHAT THIS RUN SAYS
Results will appear after the model runs.
01 / TIME & DELIVERY
Portfolio-month equivalentsA shock travels through the system.
One portfolio-month is one month of the selected wind and PM-vehicle demand. Shading marks the production shock; delivery can be affected later.
02 / MATCHED-BUDGET COMPARISON
Lower delay burden is betterWhich passage is worth opening?
Four interventions spend the same abstract budget. Inventory is bought before month 1; capacity follows a month-1 decision and qualification lag. Timing, cost conversion and gains are assumed; this is not an equally timed emergency-response or dollar-return ranking.
03 / DURATION SENSITIVITY
Cumulative delay burdenTime changes the answer.
30, 90 and 180 days are discretized as 1, 3 and 6 model months. This sweep retains the other selected inputs.
MATERIAL BALANCEInspect the physical ledger +
Initial material + external feed = deliveries + process losses + ending stocks and pipeline.
Neodymium and dysprosium are tracked separately. Both are needed to deliver the fixed technology basket; neither is used as a substitute for the other.
04 / A transparent method
Evidence underneath.
Assumptions in view.
An independent exploratory implementation, informed by Cheng, Davis and colleagues’ 2025 trade-risk study.
Deployment × material intensity
Selected wind designs and passenger-motor inputs determine annual contained neodymium and dysprosium demand. The source distinguishes direct-drive and geared permanent-magnet designs.
Read Cheng et al. (2025)A monthly delivery experiment
Three aggregate production groups split capacity across China and the rest of the world. Finite stocks, yields, lead times and qualification delay translate a shock into delivery backlog. Nine board passages explain the industry; they are not nine independently modeled capacities.
Not a reproduction of national trade-risk results
Test when relocation is feasible
The trade study locates risk at reserve owners and allows manufacturing to relocate more readily than geology. Test that assumption for one accepted NdFeB grade, with common decision, shock and delivery dates, construction, ramp and customer qualification.
Investment rankings remain conditional
05 / The evidence ledger
Follow every number
back to its origin.
Published inputs and explicit model choices are different kinds of evidence. This ledger keeps their boundaries visible.
Loading primary sources…
MODEL BOUNDARY
This first experiment covers a fixed Nd/Dy technology portfolio, not all rare earths or all magnet demand. It does not yet include Pr/Tb, grade-specific chemistry, trade routes, customer competition, substitution or recycling feed. Results do not estimate the probability of a national shortage.