Most modern wind turbines rely on powerful magnets made from rare-earth metals, primarily a material called sintered NdFeB. Its combination of high magnetic performance and low weight enables the high power density that makes permanent-magnet turbines commercially attractive.
The NdFeB supply chain remains heavily concentrated in China across mining, refining, and magnet production. The United States and the European Union are building alternative capacity, but meaningful diversification is still some years away. Choosing a turbine built around these magnets therefore locks the wind farm into that supply chain for its entire operating life. Alternative turbine designs exist that avoid rare-earth magnets, but trade off efficiency, maintenance profile, and offshore suitability.
This assessment maps the supply-chain and geopolitical risks of choosing between magnet-based and magnet-free turbine architectures, so the procurement investment can weigh lifetime risk exposure against near-term performance.
The analysis covers the high-grade sintered magnets used in the largest direct-drive wind turbines, and traces every step from raw ore and recycled feedstock through refining, alloying, sintering, coating, and shipping to final delivery at a Nordic assembly site.
This depth matters because apparent supply diversity at the final manufacturer level often collapses into a single upstream chokepoint: the same refining plants, the same coating lines, the same logistics corridors. A procurement route that looks hedged may not be hedged at all.
The findings that follow map where those chokepoints sit and which turbine architecture choices meaningfully change the risk profile.
This question establishes the scale of the exposure. If concentration is moderate, the turbine architecture choice is mainly about performance and cost. If it is extreme and concentrated at steps with no near-term substitute then choosing PM turbines is a structural bet on a single jurisdiction for the asset's operating life, not a component-level procurement decision.
Wind turbine magnets operate hot. Without dysprosium and terbium doping, they demagnetise under load. These heavy rare earths are significantly more concentrated than neodymium, and their supply can tighten independently. A scenario where neodymium supply loosens but heavy rare earths do not still leaves wind-turbine-grade magnets exposed, which is why this bottleneck is assessed separately from the broader rare-earth question.
The lock-in argument against PM turbines rests on the assumption that there is no escape route from the primary supply chain. If secondary supply of recycled magnets from end-of-life EVs, industrial motors, and early-generation turbines can scale to a material share of demand within the wind farm's operating horizon, the risk profile softens. If it cannot, the lock-in is real and long-lived.
Together, these three questions define the conditions under which Permanent Magnets turbines are the right choice. The combined answer also depends on how much the wind farm is willing to trade on energy yield, capacity factor, and maintenance cost to reduce supply-chain exposure. The same risk picture can support a PM or non-PM recommendation depending on that tolerance, which is why the results section presents the risk findings against a range of performance-tradeoff thresholds rather than a single break-even point.
The data below presents the chain at three levels of depth: the visible Tier 1 suppliers, the full industrial structure behind them, and the subset assessed in detail. This memo draws on publicly available production data and, where gaps exist, incorporates estimates based on a combination of available information.
Tier 1 supplier geography is the visible entry point into the chain, but it does not by itself show whether apparently diversified suppliers still converge on the same upstream rare-earth and processing chokepoints.
The full chain behind those Tier 1 suppliers spans 50 industries across 4 tiers, from ore extraction through refining, alloying, sintering, coating, and logistics. Each industry has a large number of companies and even larger numbers of factories. The underlying data used in this report investigates these factory level data.
| Company | Country | Overall risk | Mapped factories |
|---|---|---|---|
| The Samarium Magnet Company | SAU | 97.2% | 1 |
| Magneti Ljubljana, d.d. | SVN | 55.2% | 1 |
| Shin-Etsu Chemical Co., Ltd. | JPN | 50.2% | 3 |
| Neorem Magnets Oy | FIN | 49.5% | 1 |
| HyProMag Ltd | GBR | 48.5% | 2 |
Factory risks aggregate into supplier risks and cause industry level risk exposure. While suppliers can be changed, each of these industries is necessary for the end product. Chokepoints and geostrategic exposure in any one of them propagate uncertainty through the entire chain.
Each point is a supplier industry containing multiple companies and factories. Hover a point to see its name and scale.
Browse necessary industries in the full supply chain
Of these 50 industries, 11 were identified as the most likely chokepoints and assessed in detail for geopolitical risk exposure.
Table shows the China exposure in different industries, taking into account their deep chains.
The supply chain remains structurally concentrated around Chinese upstream processing, even where Tier 1 supplier geography appears more diverse. The mapped chain shows high dependence on Chinese rare-earth separation, refining, alloying, and powder production routes.
The main evidence is the combination of chain-level and industry-level China exposure. Direct weighted China exposure across the mapped chain is 63.0%, while Tier 1 sintered NdFeB permanent magnet manufacturing shows 72.3% China exposure. Upstream exposure is also high in Nd/Pr rare-earth metal, NdFeB alloy, and magnet-powder supply, and in rare-earth oxide separation and rare-earth metal refining.
For continuity planning, the relevant exposure is therefore the full qualified route, not only the final magnet producer. Apparent diversification at the visible supplier level does not remove strategic dependency if qualified upstream processing routes remain concentrated in the same geography.
| Supplier industry | Mapped companies | Estimated China exposure | Mapped factories |
|---|---|---|---|
| Sintered NdFeB permanent magnet manufacturer | 143 | 100.0% | 202 |
| Dy/Tb dopant supplier for higher-temperature grades | 55 | 100.0% | 70 |
| Heavy rare-earth oxide separation and heavy rare-earth metal refining | 32 | 99.9% | 50 |
| Nd/Pr rare-earth metal / NdFeB alloy / magnet-powder supplier | 55 | 99.9% | 73 |
| Recycled magnet feedstock processor | 28 | 97.2% | 29 |
| Secondary rare-earth feedstock recycler | 27 | 86.0% | 35 |
| Ionic-clay or other HREE source producer | 19 | 51.4% | 50 |
| Rare-earth concentrate producer / rare-earth ore miner | 28 | 51.4% | 33 |
| Rare-earth oxide separation and rare-earth metal refining | 31 | 50.0% | 59 |
| Heavy rare-earth feedstock producer | 26 | 46.5% | 54 |
| Unconventional rare-earth feedstock processor | 38 | 4.4% | 53 |
The most acute upstream bottlenecks are in heavy rare earths required for higher-temperature grades. Dysprosium and terbium supply remains concentrated in a limited set of source, separation, and dopant-processing industries, with especially high China exposure in the critical refining stages.
Heavy rare-earth oxide separation and heavy rare-earth metal refining show 90.3% China exposure. Dy/Tb dopant supply for higher-temperature grades shows 77.0% China exposure. Ionic-clay or other heavy rare-earth source production is also concentrated, with 51.4% China exposure.
Even if Nd/Pr exposure were reduced, high-temperature magnet grades would still face a distinct Dy/Tb bottleneck.
This bottleneck is not relieved by the diversification trajectories currently most visible in public discussion, including expanded light-rare-earth capacity outside China, additional Tier 1 magnet manufacturing, or policy frameworks framed around rare earths in general. Heavy rare earths are a separate critical path, and one that the broader "reduce rare-earth dependency" conversation does not automatically address.
| Supplier industry | Mapped companies | Estimated China exposure | Mapped factories |
|---|---|---|---|
| Dy/Tb dopant supplier for higher-temperature grades | 55 | 100.0% | 70 |
| Heavy rare-earth oxide separation and heavy rare-earth metal refining | 32 | 99.9% | 50 |
| Ionic-clay or other HREE source producer | 19 | 51.4% | 50 |
| Heavy rare-earth feedstock producer | 26 | 46.5% | 54 |
Suppliers in heavy rare-earth separation, refining, and Dy/Tb dopant supply.
Recycled feedstock cannot be counted on as an independent resilience pathway at the point of procurement commitment. Recycling routes exist, but mapped recycling and secondary feedstock processing remain highly exposed to the same geographic concentration patterns that shape the primary supply chain.
Recycled magnet feedstock processing shows 97.2% China exposure, and secondary rare-earth feedstock recycling shows 85.9% China exposure. In the current chain configuration, recycling does not function as an independent resilience pathway — it remains tied to the same concentrated downstream processing and refining capacity as primary supply.
Recycling may become strategically relevant later in the asset's operating life, particularly if non-Chinese collection, separation, and refining capacity expands. The finding here is narrower: within the window during which the CapEx decision gets made, recycling cannot be treated as a confirmed substitute for concentrated upstream primary supply. It should be tracked as a developing route rather than relied on as a mitigation.
| Supplier industry | Mapped companies | Estimated China exposure | Mapped factories |
|---|---|---|---|
| Recycled magnet feedstock processor | 28 | 97.2% | 29 |
| Secondary rare-earth feedstock recycler | 27 | 86.0% | 35 |
| Heavy rare-earth feedstock producer | 26 | 46.5% | 54 |
| Unconventional rare-earth feedstock processor | 38 | 4.4% | 53 |
The findings establish three things: the chain is structurally concentrated around Chinese upstream processing, heavy rare earths form a separate and tighter bottleneck, and recycling cannot substitute for primary supply at the point of commitment. Together these define the risk side of the CapEx decision.
The recommendation depends on how that risk is weighed against the performance the wind farm would otherwise get from PM direct-drive turbines. The comparison below draws on published OEM specifications and independent industry data; it is not primary research, but it calibrates the risk findings against the tradeoff the procurement decision actually requires.
The four turbine architectures now in serial production differ in how much rare-earth magnet material they require and in which siting contexts they are competitive. The table below summarises the procurement-relevant differences, and is meant to calibrate the risk findings rather than replace project-specific engineering assessment.
| DFIG (geared) | Medium-speed PM | Direct-drive PM | EESG (direct-drive) | |
|---|---|---|---|---|
| Generator PM content | None | ~160 kg/MW | ~650 kg/MW | None |
| Offshore / access-limited siting | Weakens at large scale | Strong compromise | Strongest | Technically viable, uncommon |
| Onshore maturity | High — mature fleet | High and growing | High | Narrower (primarily Enercon) |
| O&M profile | Gearbox + slip rings; well-understood failure modes | Gearbox present, simpler generator than DFIG | No gearbox; fewer moving parts, but large-generator interventions are serious | No gearbox and no PM; excitation system adds complexity |
| Generator rare-earth exposure | None | Moderate (NdFeB + some Dy/Tb) | High (NdFeB + Dy/Tb) | None |
| Representative platforms | SG 5.0-145, GE 6 MW onshore family | Vestas V236-15.0, Vestas EnVentus, Goldwind MSPM | SG 14-236, GE Haliade-X, Enercon E-160 EP5, Goldwind PMDD | Enercon E-138 EP3 |
Two nuances matter for reading this table. First, the performance gap between direct-drive PM and medium-speed PM at large offshore scales is narrow: there is lower O&M costs for 15 MW direct-drive in access-constrained offshore cases, but the gap was much smaller than historically claimed and sensitive to site assumptions. Second, whole-life climate impact is effectively equivalent across drivetrains (6.8 vs. 7.1 g CO₂-eq/kWh in recent prospective LCA work). The drivetrain choice is a supply-risk and bankability question.
Continuity risk sits in the full route from ore or recycled input to qualified wind-grade magnet. The performance comparison above shows that alternatives to PM direct-drive are real but not costless, and that the right choice depends on where the wind farm sits on four axes: siting context, O&M access economics, onshore bankability constraints, and tolerance for rare-earth exposure.
The following four procurement paths correspond to coherent risk-appetite profiles:
Minimum rare-earth exposure. DFIG (onshore, mature scale) or EESG direct-drive (Enercon platforms onshore). Neither uses NdFeB in the generator. Suited to buyers whose priority is eliminating the supply-risk exposure mapped in the findings, accepting in return a narrower OEM shortlist and — for EESG — a heavier generator with excitation losses.
Lowest combined rare-earth and offshore competitiveness. Medium-speed PM (Vestas V236-15.0, Vestas EnVentus, Goldwind MSPM). Around one-quarter the magnet content per MW of direct-drive PM, while retaining full-converter grid behaviour and competitive offshore LCOE. The strongest compromise for buyers who want offshore viability without the ~650 kg/MW magnet exposure.
Lowest offshore O&M burden in access-constrained sites. Direct-drive PM (Siemens Gamesa SG 14-236, GE Haliade-X). Highest rare-earth exposure, but the strongest case where major-component replacement windows are scarce and gearbox avoidance carries measurable value. Defensible when supply-chain mitigation (dual qualification, route transparency, contractual hedges) is contractually secured rather than assumed.
Maximum onshore bankability and conservative lender stance. Mature DFIG or established geared PM platforms with long operating fleets. Risk mitigation via OEM maturity and serial-defect history rather than architectural choice. Most relevant where financing structure rewards track record over performance optimisation.
Floating offshore is a cross-cutting case: the architecture choice is not pre-determined, but nacelle mass and heavy-repair strategy weigh more heavily, which tends to favour medium-speed PM in current next-generation designs. For any PM-based path, the four actions below apply regardless of which specific architecture is selected.
First, supplier qualification should be evaluated by full upstream route, not only by Tier 1 company or country of final assembly. A non-Chinese Tier 1 supplier may still depend on highly concentrated Chinese heavy rare-earth or refining inputs.
Second, heavy rare earth exposure should be treated as a distinct strategic risk. Procurement, qualification, and contingency planning should explicitly identify dependence on Dy/Tb-bearing routes for higher-temperature grades.
Third, recycling should not be relied on as a primary mitigation at the point of procurement commitment. It may support longer-term resilience, but current recycling pathways are too concentrated to materially reduce dependency at the decision moment.
Fourth, continuity planning should focus on chokepoint industries rather than only on end suppliers. The most decision-relevant exposures are in heavy rare-earth oxide separation and metal refining, Dy/Tb dopant supply, recycled feedstock processing, and Nd/Pr metal, alloy, and powder production.