Terbium & Dysprosium
From Mine to Magnet
The investment question is no longer whether more Dy/Tb can be mined, but whether separation, alloy, and magnet capacity can be financed and built outside China before 2028.
China Holds ~99% of Heavy REE Separation as Downstream Concentration Hardens
Behind 390,000t of global REO sits a much thinner Dy/Tb market, and that slice still funnels through Chinese midstream and Myanmar-linked feed before it becomes magnet metal.
China 270,000t (69%), mostly light REE from bastnäsite/monazite
USGS MCS 2026Heavy magnet feed is a thin slice of total REO tonnes
USGS MCS 2026Assumes today’s magnet chemistry and motor designs
Adamas IntelligenceMerchant heavy separation outside China only began at Lynas LAMP in 2025
CSIS 2025Shared IAC geology; Myanmar ore still exits east into Yunnan
IEA Rare Earth Pathways 2025In 2024, the single largest non-China mined Dy/Tb source, and conflict-exposed
IEA Rare Earth Pathways 2025Mine Tonnes Diversify Faster Than Separation, Metals, or Magnets
Mine supply is broad, but heavy-REE mine supply is not.
Global mine output reached 390,000t REO in 2025, of which China produced 270,000t. Terbium and dysprosium are concentrated in ion-adsorption clays and a handful of xenotime deposits, while bastnäsite and monazite, which account for most global tonnes, remain more than 95% light REE.
China and Myanmar anchor mined heavy magnet rare earth supply.
Together they supply roughly 83% of mined heavy magnet rare earths, drawing on the same weathered-granite IAC geology across Kachin and Jiangxi. Projects such as Serra Verde in Brazil (2024) and Aclara in Chile/Brazil are advancing, but commercial IAC scale and co-located separation are still centered in Southeast Asia.
Myanmar is globally material to Dy/Tb availability.
Myanmar accounted for more than 40% of mined Dy/Tb in 2024. After China banned surface heap-leaching in 2012, extraction migrated into Kachin, and concentrate still returns to Chinese separators. Myanmar has no local plant, port, or southbound road corridor.
Ion-adsorption clays are the heavy-REE foundation.
IACs supply about 80% of global heavy rare earths. Tropical weathering leaves HREEs adsorbed on near-surface kaolinite, so ammonium sulfate leaching recovers them without the blasting or flotation required for hard-rock bastnäsite and monazite.
The true chokepoint is separation, not mining.
Roughly 90% of REE separation capacity sits in China. Separating adjacent lanthanides requires hundreds of mixer-settler stages, and the cascade chemistry pioneered by Xu Guangxian in the 1970s and 1980s has compounded into proprietary know-how that China export-banned in December 2023.
Heavy rare earth processing is still concentrated in one country.
Nearly all HREE processing remains Chinese, and Dy/Tb/Y separation windows are tighter than those for Nd/Pr. Lynas’s 2025 merchant Dy oxide was the first commercial HREE separation outside China; Shin-Etsu had previously produced only small internal volumes for its own magnets.
Magnet leverage remains overwhelmingly Chinese.
China produces about 90% of permanent magnets. NdFeB originated with Sumitomo in 1982, but Chinese joint ventures later co-located strip-casting, jet-milling, and metal production with separation, creating a cost structure that Western standalone plants struggle to match.
The best ex-China midstream projects are real but still narrow.
Lynas can deliver up to 1,500t/yr of SEG/H, with Kalgoorlie providing 68kt/yr of MREC feed on a Mount Weld chain built over two decades. That is credible midstream capacity, but it is not representative of the broader project pipeline on paper.
The next bottleneck may emerge before the West fully scales.
An HREE deficit could appear by 2028 if metallization, alloying, and magnet capacity lag. Ex-China mines can add oxide faster than conversion plants absorb it, so surplus oxide still tends to route back to China.
Tightness supports a structurally higher price deck.
The consensus 2040 case sits near US$1,000/kg for Tb oxide and US$400/kg for Dy oxide. That deck is supported by EV magnets that need 2–10% Dy/Tb at 180–200°C and by direct-drive offshore wind using up to ~600 kg NdFeB per MW, while export licensing continues to slow supply response.
EVs and Direct-Drive Wind Set the Dy/Tb Curve; Humanoid Robotics Could Match Wind by 2035
China consumes roughly 70% of global NdFeB, so producer leverage and domestic demand move together. Start with the intensity columns: a segment’s share of magnet demand can understate Dy/Tb exposure when unit loadings differ by orders of magnitude.
NdFeB Demand by End-Use
| Segment | Share of NdFeB Demand | Magnet Intensity | Driver | Growth |
|---|---|---|---|---|
| EV traction motors | ~35–40% | ~1.5 kg NdFeB per vehicle (2–10% Dy/Tb by mass of magnet) | Battery-electric vehicle penetration and motor size drive the curve, and Chinese domestic EV production dominates the demand base. | Primary growth curve through 2040 under IEA STEPS and APS scenarios. |
| Direct-drive wind turbines | ~15–20% | Up to ~600 kg NdFeB per MW of direct-drive capacity | Offshore wind build-out is the main driver, and direct-drive machines now account for roughly 90% of offshore installations. | Growth tracks offshore capital-expenditure cycles and the rising direct-drive share of onshore turbines. This remains the most terbium-heavy demand segment. |
| Industrial motors & robotics | ~15% | Roughly 0.2–5 kg of NdFeB per industrial unit, and about 2–3 kg per humanoid robot. | Factory automation and emerging humanoid robot volumes, including programs such as Tesla Optimus, 1X, and Figure, set the outlook. | This is a new demand class emerging between 2026 and 2030, with potential to match wind’s Dy/Tb pull by 2035. |
| Consumer electronics & HDDs | ~15% | Grams per unit at very high volume | Smartphones, speakers, hard-disk drives, and earbuds remain the volume base, with HDDs in decline. | Demand is flat to declining as hard-disk drives lose share to solid-state storage. |
| Defense & specialized | ~5–10% | Guidance systems, electric propulsion, and F-35 avionics can require on the order of 400 kg of rare earths per aircraft. | U.S. Department of Defense stockpile targets, Navy electric-drive programs, and hypersonic guidance systems drive specialized demand. | Growth arrives in step-function expansions tied to stockpile legislation. |
China’s Share Climbs from 69% at the Mine to Near-Total Control Downstream
Western projects tend to rebuild each stage on separate sites, while Chinese incumbents already co-locate oxide, metal, and magnet production.
China Share by Value-Chain Stage
Bayan Obo Pays for LREE; Southern IAC and Myanmar Feed the Heavy Stream
Beijing can also use quotas to tighten both state-allocated tonnes and gray-market output without changing export licenses.
Chinese Producer Map
| Producer | Region | Ore type | Role |
|---|---|---|---|
| China Northern Rare Earth Group | Inner Mongolia (Baotou) | Bayan Obo bastnäsite, LREE-dominant | The largest single rare-earth producer globally, supplying LREE feed into the domestic separation and magnet chain. |
| China Rare Earth Group | Jiangxi, Guangdong, Fujian, Guangxi, Hunan | Southern IAC, HREE-dominant | Consolidated southern IAC producer formed in December 2021 by merging Minmetals, Chinalco, and Ganzhou groups. |
| Xiamen Tungsten | Fujian | IAC + magnet-grade metals | Integrated mine-to-magnet producer with a large share of Chinese magnet exports. |
| Shenghe Resources | Sichuan HQ; global stakes | Trader, not primary miner | China’s main international offtake counterparty and overseas investor |
| China Minmetals Rare Earth | Hunan + merged into China Rare Earth Group | IAC | Now part of China Rare Earth Group after 2021 consolidation |
China runs two tracks of output: state-allocated tonnes cited in WTO contexts, and unofficial or gray production. Environmental enforcement from 2020 to 2023 cut the unofficial track and pushed more demand into Myanmar imports. The April and October 2025 export-control expansions sit on top of the quota system as a second policy lever.
Myanmar Alone Accounted for 40%+ of Mined Dy/Tb in 2024
Geographic concentration at the mine and monopoly at the midstream compound each other, so a shock in Kachin or a Chinese licensing delay hits Dy/Tb magnets long before any new Western mine can respond.
China and Myanmar draw on the same weathered-granite IAC belt. Xenotime deposits such as Browns Range and Lofdal are the other HREE-rich ore class, but they have never matched IAC at commercial scale.
IEA Rare Earth Pathways 2025, p.31After China banned domestic heap-leaching in 2012, extraction shifted into Kachin. With no local separator, port, or southbound road, the only physical exit remains into Yunnan.
IEA Rare Earth Pathways 2025, p.31; CSISHREE cascades need more stages and tighter impurity windows than Nd/Pr separation. Lynas’s 2025 merchant Dy oxide was the first commercial step outside China; Shin-Etsu had previously held only small internal volumes for its own magnets.
Browns Range DFS, p.111; CSISFeed moves from Mount Weld through Kalgoorlie crack and leach into LAMP separation. That twenty-year integrated build is the benchmark for what credible midstream capacity looks like outside China.
ADL Rare Earth Imperative 2025, p.17Kalgoorlie keeps low-level radioactive cracking residues in Australia rather than Malaysia, removing a permitting failure mode that has killed other Western projects.
ADL Rare Earth Imperative 2025, p.17Demand is supported by heat-stable EV NdFeB magnets that use 2–10% Dy/Tb and by direct-drive offshore wind that can require up to ~600 kg NdFeB per MW, while export licensing continues to limit supply elasticity.
Nechalacho investor materials, p.26; Adamas IntelligenceAnnounced Projects Do Not Yet Equal an Ex-China System
Press releases can announce tonnes, but a working system still needs a physical exit route, licensed separation chemistry, and financed metals capacity arriving on the same timeline.
Kachin has no separator, deep-water port, or southbound road, so logistics only run east into Yunnan. Tonnes mined outside China still cannot reach another processor.
IEA Rare Earth Pathways 2025, p.31The export ban covers extraction and separation process IP, including the solvent-extraction cascade Western projects would otherwise copy to skip a decade of development.
CSIS China REE Export Restrictions 2025, p.5Seadrift targets both light and heavy REE separation. Outside China, commercial HREE separation at scale today is effectively limited to Lynas LAMP.
Camoin Rare Earth Call To Action 2025, p.11Browns Range xenotime carries about 10% Dy/Tb in TREO, versus under 5% for most hard-rock ores. Heavy-REE mines only work when separation is pre-committed; isolated mines idle.
Browns Range Project, p.32Lofdal and Browns Range offer lower thorium, tighter HREE grade, and fewer separation stages than monazite, which is why diversification capital concentrates there.
Namibia Critical Metals Lofdal, p.8Ex-incumbent metals, alloys, and magnets total only about 18 kt combined. The binding constraint is integrated feedstock, strip-casting, and jet-milling, not mine tonnes.
IEA Rare Earth Pathways 2025, pp.10, 48Grain-Boundary Diffusion, Tesla REE-Free Motors, and Iron Nitride Compress the 2040 Deficit Case
The 2040 deficit softens when magnets use less Dy/Tb per motor, not only when more oxide arrives from new mines, so commercially proven thrift bends the curve faster than announced substitutes that are still waiting on a production timeline.
Grain-boundary diffusion (GBD)
30–50% less Dy/Tb per magnetGBD places Dy and Tb only at grain boundaries rather than throughout the bulk alloy, so magnets keep the same coercivity and heat resistance at a lower heavy-REE loading. It is already the largest near-term thrift lever in commercial production.
Hitachi/Shin-Etsu technical papers; industry reportingTesla REE-free traction motor (planned)
Removes Nd/Pr/Dy/Tb from one OEM BOMTesla has planned a next-generation drive unit without rare-earth magnets, but the timeline and design remain undisclosed. If Tesla volume follows that path, the change would be material to the NdFeB demand curve.
Tesla 2023 Investor Day presentationIron nitride magnets (Niron)
REE-free for selected applicationsNiron in Minneapolis is developing iron nitride magnets with support from DARPA, Volvo, Stellantis, and Allison. The technology is still sub-scale today, but it is the first REE-free permanent magnet route with clear production intent.
Niron Magnetics 2024–25 disclosures; DoE funding recordsFerrite magnet substitution
No REE; ~3× lower performance/kgFerrites remain viable for appliances, low-end industrial motors, and some pumps, but not for EV traction or offshore wind. Their role is to compress NdFeB growth at the low end of the performance stack rather than replace it in high-intensity uses.
IDTechEx; AdamasDiversified Regions Need ~2× Mining, 4× Refining, and 6× Magnet Capacity Beyond Plans
Even a fully financed ex-China mine underdelivers security if refining and magnet plants remain the thin links, because oxide without a local metal and magnet path still defaults back into the incumbent system.
Ex-China and Myanmar capacity versus demand falls from mining to refining to magnets. Each stage raises capital, IP, and permitting intensity, while Chinese sites already move oxide, metal, and magnets inside one complex.
IEA Rare Earth Pathways 2025, p.47Meeting diversified-region demand still requires about 2x more mining, 4x more refining, and 6x more magnet capacity beyond current plans. An under-built magnet plant strands oxide and then the mine behind it.
IEA Rare Earth Pathways 2025, p.47Diligence should track named plants such as MP Independence, USA RE Stillwater, Noveon San Marcos, Neo Narva, SGI Vina, Star Group Daegu, Vulcan Durham, Phoenix Exeter, and JS Link Columbus. Company-level aggregates often hide stage gaps.
IEA Rare Earth Pathways 2025 + company materialsThe U.S. chain runs across Mountain Pass, HREE expansion, Independence, Stillwater, San Marcos, Durham, Exeter, and Columbus. Because the December 2023 IP ban blocks shared process tech, each stage needs its own know-how and therefore builds more slowly than an integrated incumbent.
MP Materials 10-K 2025; CSIS 2025; Camoin 2025; Adamas Oct 2025Ex-China Corridors Are Real, but Most Still Break Before Metals or Magnets
The corridors that matter already clear mine or midstream stages in places, yet finished Dy/Tb magnet security still depends on whether metal and alloy capacity lands on the same route.
U.S. Mountain Pass to Magnetics
Operating + rampMountain Pass linked to Fort Worth electrowinning, strip casting, precursor, and magnet capacity is the clearest U.S. mine-to-magnet route now in commercial ramp.
Heavy-REE feedstock, HREE separation scale, and metal-to-alloy conversion still lag the mine and magnet headlines, so the corridor is stronger on light REE magnets than on Dy/Tb depth.
Australia to Malaysia midstream
Operating + expansionLynas links Mount Weld feed through Kalgoorlie crack-and-leach into LAMP separation, which remains the strongest non-China midstream chain with a visible heavy-REE step-up already operating.
Metals, alloys, and finished magnets are not equally mature on the same route, so midstream strength still needs a separate downstream partner or build to reach magnet security.
Estonia separation to magnet
Operating, narrowNeo Silmet separation feeding Neo Narva magnets is Europe’s clearest operating separation-to-magnet pair, and for now it is almost the entire European bench that is actually running.
Adjacent European metals and alloy options such as GKN Powder Metallurgy and Lacq are still at an early facility-definition stage, so the corridor stays narrow even while both Neo plants operate.
Korea–Vietnam magnet arc
Downstream operatingOperating magnet plants across Korea and Vietnam already make this the most developed downstream arc outside China.
Upstream and midstream remain incomplete on the same arc, and both VTRE’s operating status and Star Group’s U.S. leg remain unclear, so magnet capacity still depends on imported feed.
A Non-Chinese Mine Still Usually Means a Chinese Separator
Most offtake still ends in Chinese separation. Five merchant-scale plants remain the credible ex-China alternatives; Russian capacity is isolated and excluded.
Mine-to-Separation Routing
| Mine / project | Country | Status | Where separation happens | Ex-China? |
|---|---|---|---|---|
| Mount Weld | Australia | Operating | Cracking and leaching at Kalgoorlie, then separation at LAMP in Malaysia | Ex-China |
| Mountain Pass | USA | Operating, with own separation still ramping | Historical: Shenghe Resources (China). Current: MP Mountain Pass own separation plant, commissioning 2023–24 | Ramping |
| Serra Verde | Brazil | Operating since 2024 | Reported offtake: Shenghe Resources (China). Ex-China routes via Solvay (France) and Energy Fuels White Mesa (USA) in development | Still China |
| Nechalacho | Canada | Paused / distressed | Pilot concentrate shipped to Shenghe (China). REEtec (Norway) route collapsed in 2023 | Still China |
| Browns Range | Australia | Pilot stage, with future production planned | Historical pilot: Chinese offtakers. Future target: Iluka Eneabba refinery (WA) | Pre-production |
| Aclara Penco / Carina | Chile / Brazil | Pre-production | Planned: ex-China toll processing (VAC magnet offtake, U.S. DOE support). No operating route yet | Pre-production |
| Lofdal | Namibia | Pre-feasibility | No separation partner committed yet; likely future European or U.S. toll processing | Pre-production |
| Nolans | Australia | Construction | Planned: integrated Australian mine-to-oxide, internal separation | Pre-production |
| Round Top | USA | Development | Planned: Stillwater, Oklahoma separation. Not yet operating | Pre-production |
| Indian Rare Earths Ltd | India | Operating | Domestic state-controlled separation | Ex-China |
Merchant-Scale Ex-China Separation Operating Today
| Plant | Country | Scale | Notes |
|---|---|---|---|
| Lynas LAMP | Malaysia | Industrial | The only non-Chinese separation plant at true industrial scale. Expanding to add SEG/H heavy-REE output through 2025–26. |
| Neo Performance Silmet | Estonia | Small | An existing small separation plant, and the only operating European site producing separated rare earths today. |
| Solvay La Rochelle | France | Small, specialised | Legacy European separation with small volumes, mostly oriented downstream. |
| MP Materials Mountain Pass | USA | Ramping | MP’s own separation has been commissioning since 2023–24. A DoD $400M equity stake and 10-year magnet offtake were committed in 2025. |
| Indian Rare Earths Ltd | India | Domestic | State-controlled capacity that serves Indian domestic needs rather than the global merchant market. |
Magnet Recycling Could Outsize Most Ex-China Mines Now in Development
Three feedstock classes matter most. Manufacturing swarf is already commercial, end-of-life EV motors begin to ramp around 2028–30, and HDD or consumer reclaim follows the Japanese template. Models that count only primary mines will understate the supply balance outside China if the upper recycling case materializes.
Recycling Routes and Named Operators
NdFeB grinding generates 20–30% swarf. Closed-loop recovery back into magnets is already mature in Japan and at selected U.S. sites, and it remains the lowest-cost recovery route per kilogram.
First-generation EVs begin to retire around 2028–30, unlocking Dy/Tb-bearing NdFeB from traction motors. Hydrometallurgy and direct-remelt routes are commercializing ahead of that retirement wave.
JOGMEC has coordinated HDD magnet reclaim since the mid-2010s. Volumes are small, but the process template is proven and useful for U.S. and EU policy design.
Financing Closure, Not Announcements, Sets the Realistic 2028 Dy Supply Base
The table includes only financing stacks with a disclosed sponsor, instrument, and size. Projects that do not appear here still lack a closed capital path into 2028 tonnes.
Closed or Committed Ex-China Financing
| Project | Sponsor | Instrument | Size | Stage |
|---|---|---|---|---|
| MP Materials (Mountain Pass + Fort Worth) | US Department of Defense | Equity + 10-year magnet offtake + price floor | $400M equity (mid-2025) | Ramping production |
| MP Materials HREE separation expansion | US Department of War (ex-DoD) | Loan | $150M (Aug 2025) | Construction |
| Lynas Rare Earths (Kalgoorlie and Texas HREE plant) | US DoD DPA Title III + EXIM Bank + Australian EFA | Grants + export credit | $258M Lynas + separate Texas award | Kalgoorlie commissioned; Texas construction |
| USA Rare Earth (Round Top + Stillwater) | Private + DoD development | Private equity + strategic | Undisclosed; magnet plant operational Apr 2026 | Magnet production started; upstream in development |
| Arafura Nolans | EXIM Bank + Export Finance Australia + EU CRMA | Senior debt package | ~A$1.1B committed | Construction |
| Iluka Eneabba refinery | Australian Government | Non-recourse loan | A$1.25B | Under construction; commissioning slipping toward 2027 |
| Aclara Penco / Carina | US DOE + VAC (Germany) offtake | Grants + offtake | Undisclosed size | Pre-production |
| Northern Minerals Browns Range | Australian Critical Minerals Facility + Iluka offtake | Debt + offtake | A$200M+ senior debt | Pilot; ramp conditional on Eneabba |
| Japanese magnet / recycling industrial policy | METI / JOGMEC | Equity stakes + stockpile purchases + R&D grants | Multi-decade, aggregate ~¥2T | Continuous since 2010 Senkaku embargo |
Vietnam Is a Platform Design Problem Because Geology Alone Does Not Create Optionality
Countries that look geologically similar to southern China still trade at a discount until they can clear a commercial separation path and keep concentrate from defaulting east into Chinese midstream.
Resource is addressable, not bankable by itself
Northern Vietnam shares weathered-granite ion-adsorption geology with Jiangxi and Kachin, and revised estimates cite roughly 3.5 million tonnes of TREO. The missing piece is commercial heavy-REE separation, because reserve tonnes do not become separated oxide without a process route.
Execution package is the underwriting object
Strategic capital prices license clarity, process pathway, site utilities, governance, and offtake visibility ahead of headline geology alone.
Buyer set is industrial and policy-backed
The likely counterparties are processors, magnet makers, offtakers, and policy-backed capital seeking non-China optionality through a structured entry.
Tracked Rare Earth Assets Still Cluster Processing Inside China
Mines appear across more countries than separators and magnet plants do, which is why a diversified asset map can still hide a single-jurisdiction midstream concentration.
Global Rare Earth Asset Map
Loading tracked asset inventory…
Diversification Structurally Fails at Separation, Not Mining
Mining can look plural across more countries, but the concentration tightens at every processing step until magnets again sit inside a near-single-jurisdiction system.
Stage 1: Mining: Ore Extraction
China: 69%HighIon-adsorption clays and xenotime host the economic terbium and dysprosium grades. Bastnäsite and monazite deposits such as Mountain Pass, Mount Weld, and Bayan Obo remain more than 95% light rare earths. Commercial IAC production is concentrated on weathered granite in southern China and Kachin’s Myanmar belt, where ammonium sulfate leaching recovers the metals without blasting or flotation. After China banned domestic surface heap-leach IAC in 2012, Chinese demand migrated into Kachin, and the concentrate still returns to Chinese separators. Serra Verde in Brazil is producing and Aclara’s Chile and Brazil projects are in development, but commercial scale and co-located processing remain centered in Southeast Asia.
No tracked assets are available for this stage in the current inventory.
Stage 2: Separation: Solvent Extraction, The Chokepoint
China: 90%ChokepointSeparating adjacent lanthanides requires hundreds of mixer-settler stages. Xu Guangxian’s cascade chemistry from the 1970s and 1980s, compounded over forty years of operating know-how, is the real moat, and China banned the export of that process IP in December 2023. Heavy-REE impurity windows are tighter than those for neodymium and praseodymium. Refining outside China covers roughly 20% of diversified-region demand today, and only about 25% even after planned expansions through 2035. The credible response sites are Lynas LAMP and Kalgoorlie, MP Mountain Pass, Neo Silmet, and Solvay La Rochelle, together with a next wave of U.S. and Australian projects that are still building from scratch.
No tracked assets are available for this stage in the current inventory.
Stage 3: Metal & Alloy Production: Oxide to Metal Reduction
China: 99%CriticalConverting oxide to metal requires anhydrous halide feedstock and molten-salt electrowinning. Chinese sites typically co-locate that step with separation, while Western projects often split sites and logistics. Browns Range concentrate carries roughly 10% Dy/Tb in TREO, dense enough to support dedicated heavy-REE metallization. MP Independence and Phoenix Exeter, which starts at about 200 tonnes per year of metals, are early U.S. steps. Industry already flags an HREE deficit risk by 2028 if conversion lags mine oxide growth, because new Western tonnes still tend to route back to China for metal.
No tracked assets are available for this stage in the current inventory.
Stage 4: Magnet Manufacturing: NdFeB Permanent Magnets
China: 92%Very HighAdding 2–10% dysprosium or terbium to NdFeB raises usable temperature from roughly 150°C to more than 200°C, which matters for EV traction magnets of about 1.5 kg per vehicle and for direct-drive offshore wind of up to roughly 600 kg per MW. NdFeB originated at Sumitomo in 1982, after which Chinese joint ventures co-located strip-casting and jet-milling with separation and metal production. Magnet capacity outside China covers only about 14% of diversified-region demand today, and remains below 20% even after planned expansions. The plants that matter for diligence are Shin-Etsu, Neo Narva, SGI Vina, Star Group Daegu, Noveon San Marcos, MP Independence, USA Rare Earth Stillwater, Vulcan Durham, and JS Link Columbus.
No tracked assets are available for this stage in the current inventory.
Tb Peaked at $2,050/kg in 2022; Dy Undersupply Still Points to 1,800t/yr by 2040
Price spikes show how quickly policy and logistics can move thin heavy-REE markets, while the 2040 gap shows how far today’s magnet chemistry still stretches Dy/Tb demand relative to financed supply.
Terbium (Tb)
Atomic number 65. Average price of $1,010/kg in 2025.
Dysprosium (Dy)
Atomic number 66. Average price of $239/kg in 2025.
Beijing Calibrated Export Pressure Through 2025 as Licensing Power Stayed Downstream
China kept the ability to tighten Dy/Tb flows through licensing while easing the blunt edge of a permanent ban, so buyers faced rationing risk without a trigger that would force a full Western midstream build overnight.
2025–2026 policy and market sequence
Northern Minerals publishes Browns Range DFS (8% of global Dy/Tb)
China expands rare-earth export controls to all heavy rare earth elements
China suspends the early-October HREE controls for one year while license pressure remains
MP Materials commences NdFeB magnet manufacturing in Fort Worth
USA Rare Earth begins magnet production (600t/yr capacity)
Myanmar Is the Single-Country Risk Inside the Single-Country Concentration
Only three scenarios are modeled. The base case is rolling stoppages, as in February 2025 when exports fell to 311 tonnes against a normal month of roughly 4,000–5,000 tonnes. The severe case removes more than 40% of mined Dy/Tb for twelve months or longer before separation capacity outside China can absorb redirected feed.
Partial disruption (rolling)
Arafura’s September 2025 materials show Myanmar-to-China exports crashing to 311 tonnes in February 2025 from a typical monthly range of about 4,000–5,000 tonnes, then recovering to 4,727 tonnes by April and 5,435 tonnes by May. That pattern of seasonal or conflict-driven stoppages followed by rebounds is the baseline scenario, not the worst case.
Medium disruption (territorial flip)
Kachin Independence Army forces have gained ground in Kachin since 2023. A territorial reversal, whether the junta retakes ground, the KIA loses its taxation base, or inter-militia conflict destabilizes extraction, would halt Chinese offtake through Yunnan for multiple quarters. Dy/Tb oxide prices could spike two to four times during such a window.
Severe disruption (prolonged conflict)
A full collapse of the Kachin extraction economy would push mine projects outside China onto the supply path as the marginal Dy/Tb source. Separation capacity outside China still could not absorb redirected volume until about 2028 at the earliest.
Hundreds of Mixer-Settler Stages and 40 Years of Process IP Keep Separation the Moat
The midstream moat is not one plant or one metal grade, but chemistry know-how, wastewater permitting, integrated cost structure, and export licensing stacked on top of each other.
40-year solvent-extraction head start
Xu Guangxian at Peking University optimized heavy-REE solvent-extraction cascades in the 1970s and 1980s, and China has compounded that process base continuously ever since.
Hundreds of mixer-settler stages
Adjacent heavy lanthanides are chemically near-identical. Separating dysprosium and terbium therefore requires long cascades of organic solvent and hydrochloric acid stages that are difficult to replicate without accumulated operating know-how.
5–10 year Western permitting for acidic wastewater
Solvent extraction produces acidic wastewater, and monazite routes also leave thorium residues. China accepted environmental costs that Western jurisdictions have generally refused.
Integrated mine-to-magnet cost structure
Chinese complexes co-locate mine, separation, metal, and magnet production. Standalone Western plants therefore carry logistics and working-capital costs that integrated sites avoid.
Export controls on oxides and metals, not ore
The April 2025 controls on terbium, dysprosium, samarium, gadolinium, scandium, and yttrium target separated products rather than ore. Mining can leave China, but conversion is forced to remain inside the licensing perimeter for export.
Japan’s 15-Year Head Start Remains the Only Proven Ex-China Playbook
Fifteen years after Senkaku, Japan still offers the clearest proof that financing, magnet thrift, recycling, and stockpiles can be scaled, while primary HREE mining and metal conversion remain the unfinished part of the playbook.
Japan Sequence Since 2010
China embargoes REE exports to Japan during Senkaku Islands dispute
The embargo was the forcing event. Japan became the first major economy to treat rare-earth dependence as a national-security problem.
JOGMEC + Sojitz rescue-finance Lynas with $250M
Created the first non-Chinese separation capacity at scale (LAMP Malaysia). Every subsequent ex-China diversification plan traces to this deal.
JOGMEC begins Vietnam rare-earth technical cooperation (Dong Pao pilot)
Early template for Japanese-Vietnamese upstream partnership; foreshadows Shin-Etsu Vietnamese magnet build-out.
Shin-Etsu, Hitachi, TDK implement grain-boundary diffusion (GBD) at commercial scale
Grain-boundary diffusion cut Dy/Tb content by 30–50% per magnet and remains the single largest piece of Japanese defensive technology against Chinese export leverage.
Japan builds stockpile to ~6 months of critical REE consumption
The stockpile was designed as a buffer against a repeat of the 2010 event. The U.S. Department of Defense is now following the same model under 2023–25 legislation.
Shin-Etsu magnet plant operational in Hai Phong, Vietnam; Hitachi continues Japanese magnet volume under Proterial rename
Japanese magnet incumbents have kept their share through 40 years of Chinese cost competition and the post-2010 diversification drive.
Figures Prefer USGS, IEA, CSIS, and Company Filings, Though Downstream Shares Still Vary
Government series and company filings set the floor for mine and trade figures, while downstream magnet and metal shares still leave more room for methodology noise across publishers.
Government statistics (USGS, IEA), company filings, exchange data
Industry research (Adamas, IDTechEx), trade association data
Academic estimates, analyst commentary, industry consensus
- USGS Mineral Commodity Summaries 2026
- IEA Critical Minerals Outlook 2025
- CSIS Critical Minerals Security Program 2025
- ADL Rare Earth Imperative 2025
- Browns Range DFS — Northern Minerals
- Adamas Intelligence — REE Market Outlook
- IDTechEx — Permanent Magnets 2024-2034
- MP Materials 10-K 2025
- Iluka Resources ASX Filings
- Northern Minerals ASX Filings