On October 26, 2026, HyProMag USA announced the results of a pre-feasibility study (PFS) on expanding its U.S. rare earth magnet manufacturing business. The concept is to build a four-plant network by 2034, raising annual finished-magnet capacity to 4,192 tonnes, four times the planned capacity of the first plant once completed. However, this is not an announcement that total U.S. magnet production capacity will quadruple, and the expansion investment and start of mass production have not been formally decided. The company aims to start finishing imported magnet blocks in 2027, then establish an integrated recycling system in 2028 that produces finished products from scrap magnets collected in the U.S. A closer look at this phased plan reveals both the technology behind the large-scale expansion concept and the challenges to realizing it.
Quadrupling finished-magnet capacity: 4,192 tonnes a year planned for 2034
The first plant, planned for North Lake, Texas, is still in the detailed design stage. The baseline for the company's "four times" claim is the 1,048 tonnes per year of neodymium-iron-boron (NdFeB) magnet manufacturing capacity this plant is expected to have when completed. Under the new concept, plants would be added in South Carolina, Nevada, and Texas, for a total of four production sites. As of June 2026, the company had only presented design studies for additional facilities and expansion targets. This time, after completing the study with PegasusTSI and BBA USA, it presented specific facility layouts and profitability by feedstock type.
The company expects total annual sellable products and materials in 2034 to reach 6,000 tonnes. However, that figure includes not only finished magnets but also items such as the alloy powder used in magnet production. The breakdown in the October 26 announcement is as follows.
| Product/material | Planned annual output in 2034 |
|---|---|
| Finished NdFeB magnets | 4,192 tonnes |
| NdFeB alloy powder | 1,305 tonnes |
| Other | 503 tonnes |
| Total | 6,000 tonnes |
The 4,192 tonnes of finished magnets is exactly four times the first plant's completed capacity of 1,048 tonnes. The 6,000 tonnes, on the other hand, combines finished magnets, alloy powder, and other materials, and not all of it would ship as finished magnets. All figures are also planned values for 2034, not current production results. Assessing finished-magnet supply capacity requires distinguishing it from total output that includes powder and other materials.
The three new plants are planned to produce sintered magnet blocks before final processing. Subsequent cutting and finishing would be handled at a centralized hub inside the second Texas plant, which would also consolidate quality assurance and customer qualification work. In other words, the centralized finishing hub is not a standalone fifth plant but a facility for turning magnet blocks sent from multiple manufacturing plants into customer-ready products.
In this configuration, what matters is not only increasing sintered block output at each plant, but also whether the central hub can process those blocks without delays and ship products that meet customers' quality standards. Even with more manufacturing plants, finished-magnet supply capacity will not expand as planned if finishing and quality qualification cannot keep pace. The published figures alone do not reveal how much cost centralization would save or how long customer qualification would take.
HPMS turns scrap magnets back into powder with hydrogen, but finished products need several more steps
The core technology behind HyProMag USA's business is HPMS (Hydrogen Processing of Magnet Scrap), a hydrogen-based scrap magnet processing technology developed at the University of Birmingham. Hydrogen is applied to NdFeB scrap magnets to break them down and demagnetize them, recovering the alloy as powder. It returns magnets used in earlier manufacturing processes to a state where they can again serve as raw material for magnets, extracting useful material from waste.
However, recovering alloy powder through HPMS does not by itself yield finished magnets. The recovered powder must be further refined, shaped, and sintered. Cutting of the sintered blocks, surface coating, and magnetization follow. Only when a product meets customer-specified requirements such as dimensional accuracy, surface treatment, and magnetic properties can it ship as a finished magnet. The amount of alloy powder recoverable from scrap magnets and the amount of finished magnets ultimately sellable must therefore be evaluated separately.
The separate listing of 4,192 tonnes of finished magnets and 1,305 tonnes of alloy powder in the 2034 plan reflects these differences in manufacturing steps. Alloy powder can itself be sold as a material, but a company that buys it needs to process it further to use it in finished magnets. Even if total sales reach 6,000 tonnes, that does not mean 6,000 tonnes of finished magnets can be supplied. This distinction also matters when assessing which manufacturing processes will be established in the U.S.
A pioneering facility for commercializing HPMS has already begun operating in the UK. According to the University of Birmingham's announcement on January 15, 2026, the magnet recycling facility opened in Tyseley, UK, can process more than 400 kg of material per batch. Its processing capacity is said to be 100 tonnes a year on a single shift and more than 300 tonnes a year with multiple shifts. These are, however, the facility's assumed processing capacities, not results from a full year of production. The UK facility is a precedent for deploying HPMS at industrial scale, but it does not guarantee that the four U.S. plants will start up in the same way.
Realizing the expansion concept will take more than expanding HPMS recovery capacity. It requires both a feedstock supply network that reliably collects used magnets and processing equipment that makes customer-spec finished magnets from recovered powder. Commercialization also depends on keeping each process running continuously while maintaining quality.
Recycled feedstock only, or new alloy too? The PFS shows profitability
The PFS compares investment and profitability for two cases: using only recycled feedstock, and installing equipment that can also use newly produced alloy feedstock. The comparison covers the three new plants other than the first North Lake plant, plus the centralized finishing hub, and does not show the investment for all four plants. The main published estimates are as follows.
| PFS comparison item | Recycled feedstock only | Also using new alloy feedstock |
|---|---|---|
| Capital expenditure | $623 million | $656 million |
| After-tax net present value (NPV) | $2.062 billion | $1.854 billion |
| Internal rate of return (IRR) | 36% | 31% |
| AISC of first expansion plant in 2030 | $37.4/kg | $41.6/kg |
Installing equipment that can also handle new alloy raises capital expenditure by $33 million compared with the recycled-only case. After-tax NPV is $208 million lower, and IRR falls from 36% to 31%. AISC for the first expansion plant in 2030 is also higher in the combined case, at $41.6/kg versus $37.4/kg for recycled only. AISC is a measure of total cost including sustaining capital, and in this estimate, supporting new alloy increases both the investment burden and the unit cost.
Even so, equipment that supports new alloy has the advantage of widening the range of feedstock options. If newly produced alloy can be used in addition to material recovered from used magnets, other sources can be considered if the supply of recycled feedstock falls short. However, the difference in profitability between the two cases may not stem entirely from feedstock flexibility. The PFS compares business models built on assumptions such as product mix and selling prices, and each condition may affect the results. Moreover, being able to use new alloy does not mean all feedstock can be sourced from outside China, or that the necessary procurement contracts have been signed.
NPV converts expected future revenues and expenditures into present value, and is neither actual profit nor confirmed order value. The company's estimates use 2026 as the base year, a 7% real discount rate, and a U.S. federal tax rate of 21%. State taxes are not included, nor are federal or state government support or revenue related to carbon pricing. The estimates also assume an average 2030 selling price of $100.5/kg, which is likewise a forecast based on future market conditions. Even for an estimate that does not factor in government support, whether the assumed selling prices, utilization rates, and manufacturing costs are actually achieved is a separate matter.
From finishing in 2027 to integrated U.S. recycling in 2028
HyProMag USA plans to launch the business in phases rather than performing everything from scrap magnet collection and recycling through finished-product manufacturing in the U.S. from the start. Under the plan announced in July 2026, it will first bring finishing equipment for imported magnet blocks online in the first half of 2027. In this initial stage, it will receive up to 20 tonnes of magnet blocks from the UK and Germany and conduct trial operation of processing and finishing in the U.S. Integrated recycling using scrap magnets collected in the U.S. as feedstock comes in the next stage.
This approach allows magnet processing and finishing equipment to start up before scrap magnet collection facilities and recycling processes are ready. On the other hand, the magnet blocks used in the initial stage are imported, so the supply system will not be completed entirely within the U.S., from collection through manufacturing. The start of finishing operations in 2027 and the integrated recycling planned for 2028 must be evaluated as separate milestones.
As of July 2026, the target for starting integrated recycling was the second quarter of 2028, but the October announcement changed the wording to mid-2028. Initial annual production capacity is expected to be 678 tonnes, combining 400 tonnes of finished magnets and 278 tonnes of byproducts. The plan is to expand capacity step by step from there, so the plant will not be able to produce the 1,048 tonnes a year that North Lake Plant 1 is expected to have at completion immediately after operations begin. Plant completion, equipment commissioning, and the transition to stable mass production are separate stages.
Work is also under way on equipment for feedstock procurement. In an October 2026 announcement, HyProMag USA said it had deployed a third Inserma hard disk drive (HDD) separation unit as a mobile system. Rather than only bringing HDDs to fixed facilities as before, the initiative allows the unit to be taken to customer sites for processing. Revenue from selling the recovered printed circuit boards is said to offset part of the cost of acquiring and storing magnet-bearing feedstock.
The company expects feedstock that is already contracted or in storage to support initial trial operation and the production ramp-up. However, this does not mean it has secured all the feedstock needed to operate the four plants planned for 2034. As the business scales, it will need to increase scrap magnet collection volumes and arrange contracts for continuous supply. Securing material of stable quality will matter as much as volume.
Meanwhile, the business plan for the leading North Lake Plant 1 has also changed. The June 2026 announcement put capital expenditure at $152 million, but in October it rose to $167 million. NPV based on forecast prices and other assumptions also fell from $797 million to $723 million. The company cited schedule delays and revisions to assumptions about construction costs and operating conditions. In June it had said it was in financing discussions with financial institutions, but that was not an announcement that financing had been completed.
The first things to watch are the trial operation of finishing equipment planned for the first half of 2027 and progress toward obtaining customer quality qualification. After that, the focus will be on whether the company can secure the necessary feedstock contracts and funding and move in 2028 to an integrated recycling system that makes finished magnets from scrap magnets collected in the U.S. Achieving the 2034 target of 4,192 tonnes a year will require putting everything in place: recovered feedstock supply, manufacturing capacity at each plant, processing capacity at the centralized finishing hub, and a shipping system that includes customer qualification. This PFS shows the profitability of the large-scale expansion concept, but whether U.S. rare earth magnet supply capacity actually expands will be judged by the track record of facility start-ups and commercial production still to come.