Illustrative: a heated billet being extruded under press during projectile case production at Scranton Army Ammunition Plant, 15 May 2024. Shown for the metal-parts stage of the munitions chain; this is a steel projectile-case line, not a tungsten refining facility. Photo: US Army, Scranton Army Ammunition Plant / DVIDS, public domain.

Washington has put $450 million into a chemical stage of tungsten refining, and the fragment is where it lands

The United States Department of War announced on 14 September 2026 a $450 million redeemable preferred equity investment in The Elmet Group, made through the Industrial Base Analysis and Sustainment programme by the Office of the Assistant Secretary of War for Industrial Base Policy with the Economic Defense Unit. Part of it will establish what the department describes as the only independent ammonium paratungstate facility in North America. Ammonium paratungstate is a chemical intermediate, not a weapon, and it sits upstream of every tungsten fragment, penetrator and fragmentation sleeve the United States builds.

Technical Summary

The department states that The Elmet Group is the only United States owned, fully integrated producer of tungsten and molybdenum materials and components, and that it supports more than 100 Department of War programmes including the F-35, Patriot PAC-3, Trident D5, the Next Generation Interceptor, the Precision Strike Missile and the Virginia and Columbia class submarines. Tungsten is described in the same release as used in critical components across missiles, munitions, aerospace, propulsion, naval, undersea, electronics and industrial manufacturing. The department puts Chinese control of global supply at an estimated 85 per cent for tungsten and 40 per cent for molybdenum. For comparison at the mining layer, the United States Geological Survey gives Chinese mine production for 2025 as 67,000 tonnes of an 85,000 tonne world total, or about 79 per cent, and records no commercial tungsten mining in the United States since 2015. The department’s 85 per cent figure is stated for global supply rather than for mine output alone.

The company’s own filing puts numbers and a map on the money. More than $165 million goes to the existing plants at Lewiston, Maine, Coldwater, Michigan and Euclid, Ohio, to expand tungsten heavy alloy and powder output for applications the company lists as Patriot, Phalanx, JDAM, F-35 and rocket programmes. Approximately $150 million goes to the Springer Tungsten Complex at Imlay, Nevada, operated by Blue Moon Metals, where Elmet intends to form and operate a majority-owned joint venture with Blue Moon Metals and EQ Resources to restart and expand ammonium paratungstate conversion capacity. Up to a further $150 million is earmarked for partnerships including the Mt Carbine mine in Australia and the Barruecopardo mine in Spain. The chemical step being funded is therefore in Nevada, not Maine: the Maine, Michigan and Ohio money is powder, pressing, sintering and finished components. The Springer mine and mill are expected to restart in the fourth quarter of 2027 and the joint venture’s ammonium paratungstate conversion facility to begin operations in the second half of 2028.

Two further terms matter. The equity is structured as an initial $200 million drawdown at closing, with the department receiving redeemable preferred equity, warrants representing up to 19.9 per cent of common stock on a post-transaction basis, one independent director and one non-voting board observer. Separately and on the same day, Elmet Technologies took an indefinite delivery, indefinite quantity contract from the Defense Logistics Agency with a ceiling of up to $2 billion and a guaranteed funded commitment of $150 million, covering tungsten ores, concentrates and sodium tungstate for DLA Strategic Materials, on a five-year base ordering period to 30 August 2031 with a two-year option to 30 August 2033. Deliveries into the National Defense Stockpile are deliberately deferred until new mining and processing capacity comes online, so that stockpile replenishment does not draw on the material existing American manufacturers are already consuming.

The ordnance relevance runs through tungsten heavy alloy, a sintered composite of tungsten powder with a ductile binder. Elmet publishes three relevant defence lines: kinetic energy penetrators, blast fragmentation sleeves, and preformed fragmentation parts in sphere, ball and cube form. Its published fragment table gives a 6 mm cube at 3.78 grams, a 7 mm cube at 5.88 grams, an 8.13 mm sphere at 4.9259 grams and a 5.59 mm sphere at 1.6007 grams. Those figures compute to approximately 17.5 grams per cubic centimetre for the spheres and the 6 mm cube, and about 17.1 for the 7 mm cube, a spread consistent with the lower classes of the ASTM B777 tungsten heavy alloy specification. Against a steel fragment of identical geometry, at roughly 7.85 grams per cubic centimetre, that is about 2.2 times the mass for the same presented area.

A fragment 2.2 times heavier for the same frontal area sheds velocity more slowly. The same charge therefore throws lethal fragments further, which makes a refining decision in Nevada a quantity-distance question on every range that fires the resulting nature. ISC assessment derived from Elmet published fragment dimensions and masses

Analysis of Effects

The physics is why this is a weapons story rather than a mining one. A fragment decelerates in air in proportion to its drag area divided by its mass. Increase density while holding the shape, and the ratio falls, so the fragment retains velocity over a longer flight. For a warhead designer that buys either greater effect at a given radius or the same effect from a smaller, lighter fragment array, which in turn frees mass for fill or for guidance. It is also why tungsten is the standard substitute in kinetic energy penetrators where depleted uranium is politically or legally unavailable: density and hardness are the properties doing the work, and very few materials offer both.

The binder choice matters to anyone who has to work at the far end. Tungsten heavy alloy is not pure tungsten, and Elmet publishes both low magnetic and non-magnetic variants alongside conventional ones. That is a supplier catalogue entry rather than a doctrinal statement, but the implication for post-incident work is direct. Fragment recovery and site exploitation lean heavily on magnetic search where the fragment population is steel. A tungsten fragment population with a non-magnetic binder does not respond the same way, and a magnetic sweep that returns a clean result over such a site has established less than the operator might assume. The same point applies in reverse to fragment identification: density and dimension become the diagnostic rather than magnetic response.

Personnel and Safety Considerations

The material itself is inert, and nothing about the refining stage raises an explosives safety question on its own. The exposure appears at assembly and downstream of it. Fragment throw is treated separately from blast overpressure in the quantity-distance frameworks that govern ammunition storage and range safety, in NATO through the Allied Ammunition Storage and Transport Publication series and in United States service through the Defense Explosives Safety Regulation. Those frameworks are calibrated against the fragment populations actually held. A nature whose fragment array is tungsten rather than steel presents a different hazardous fragment distance for the same net explosive quantity, because the governing variable is how far a fragment above the lethality threshold travels, not how much explosive drove it. Any programme that substitutes tungsten into an existing fragmenting design therefore inherits a safety case question as well as a lethality improvement, and it is a question that has to be answered by test rather than by inheritance from the steel-fragment predecessor.

There is also a compliance clock running well ahead of the plant. Defense Federal Acquisition Regulation Supplement 225.7018-2(d) restricts tungsten metal powder and tungsten heavy alloy from covered countries. Through 31 December 2026 that restriction reaches atomization, calcination and reduction into powder, final consolidation of non-melt derived metal powders, and all subsequent phases of production of tungsten metal powder, tungsten heavy alloy or any finished or semi-finished component containing tungsten heavy alloy. From 1 January 2027 it extends back to the mining or production of tungsten ore or feedstock, recycled material included. The Springer conversion facility is not expected to operate until the second half of 2028. Whatever the department has bought at the chemical step, it does not arrive in time for the step-up, and the roughly eighteen months in between have to be covered by feedstock and powder that is already non-covered. That is the near-term reason the stockpile contract and the Australian and Spanish feed arrangements matter, rather than the commissioning date.

Data Gaps

Six gaps are recorded, and they are narrower than the department’s own release alone would suggest, because the company filing published the same day closes several of them. First, design capacity in tonnes per year for the Springer conversion facility is not published, so the size of the capability being bought cannot be assessed against United States demand. Second, the redemption terms and milestone conditions attaching to the preferred equity are unstated; the headline structure is published but the conditions under which the government exits, or does not, are not. Third, the department’s release names missiles, aircraft and submarines and no fragmenting warhead line; the company filing names Patriot, Phalanx, JDAM, F-35 and rocket programmes, which reaches munitions but still identifies no specific fragmenting nature by designation. Fourth, the Elmet fragment dimensional data used here is drawn from a company capability page dated 22 March 2022 and was not restated in the 2026 announcement, so current sizes and classes may differ. Fifth, no United States annual tungsten demand figure for defence, and no current import dependency percentage beyond the stated Chinese share of global supply, is published. Sixth, the basis of the department’s 85 per cent figure is not given, and it cannot be reconciled from the release against the mining-layer figure recorded by the United States Geological Survey.

Editorial responsibility. This analysis is researched, written and published under the editorial responsibility of Steve Sawyers, MIExpE VR. Every technical claim is sourced to named open-source material and rated for reliability and accuracy under NATO STANAG 2022. ISC uses its own domain-trained research and drafting system under human direction; the analysis, the judgements and the accountability for them are ours. Corrections and updates welcome: if you hold open-source data that refines or corrects any parameter here, write to [email protected] citing the specific claim and your source. Verified corrections are incorporated and credited in the revision history. Full editorial and AI policy.