President Trump told a room of roughly 200 mining executives at the State Department on Friday that America is reclaiming its rightful place as the minerals superpower of the world. The announcement behind the line was about $3 billion in federal money spread across critical minerals and battery projects: a $1.4 billion Office of Strategic Capital loan to Sila Nanotechnologies, $400 million for scandium production in Australia, $150 million with the Minnesota magnet firm Niron, and $180 million in workforce and trade grants. It follows a $10 billion EXIM facility for a domestic strategic stockpile and roughly $14.8 billion in letters of interest issued over the past year.
The commentary that followed focused on the gap between the rhetoric and the geology, which is the easy criticism and the least interesting one. American mine permitting is slow, the deposits that matter are years from production, and no amount of federal enthusiasm compresses a decade of environmental review into a presidential term. All true, and all beside the point.
The bottleneck was never the mine
The United States is already the second-largest producer of mined rare earth oxides on the planet. What it does not have is the ability to turn those oxides into anything. Concentrate leaves the country, gets separated and refined abroad, and comes back as finished product. China holds better than 90 percent of global separation and magnet manufacturing capacity, and that number, not the tonnage pulled out of the ground, is the dependency.
Separation plants and magnet factories are three-to-five-year builds. That is the actual reason the current push is not fantasy. MP Materials began commercial magnet production at its Fort Worth facility in 2025. Its second campus in Northlake, Texas is a $1.25 billion project that adds around 7,000 tonnes of annual magnet capacity for a combined total near 10,000 tonnes, commissioning in 2028. Heavy rare earth separation is being commissioned at Mountain Pass. Energy Fuels is running light rare earth separation at White Mesa and pushing into dysprosium and terbium. e-VAC is building in South Carolina. This is construction, not a press release.
The financing structure is more durable than most of the analysis credits. A ten-year Pentagon price floor with offtake on the full output of the Northlake plant, a Defense Department equity position in MP, direct EXIM loans and a funded stockpile are not executive orders that a successor revokes with a signature. They are contracts. Whoever takes office in January 2029 inherits a partially built industrial base with federal money already spent and delivery obligations attached.
The exposure nobody is pricing
Here is what the 2028 timeline actually rests on, and it is not permitting or capital.
A rare earth separation plant is a very long train of solvent extraction mixer-settlers. A magnet plant needs strip casting furnaces, hydrogen decrepitation equipment, jet mills, aligning presses and vacuum sintering furnaces. The overwhelming majority of that equipment is designed and built in China, by firms that have spent thirty years refining it against the only operating fleet at scale. There is no meaningful Western supplier base for it, because there was no Western demand for it.
Beijing understands this precisely. In June it placed MP Materials and USA Rare Earth on export control lists, and the target was not the concentrate. It was access to processing equipment and technology. The instrument aims directly at the construction phase of the American buildout, which is exactly where the American buildout currently sits.
This is the semiconductor equipment playbook run in reverse. Washington spent five years demonstrating that the way to stall a national industrial program is not to embargo the finished product, but to withhold the machines that make it and the engineers who know how to commission them. China took the lesson. A magnet plant with the building finished, the workforce hired and no sintering furnaces on the dock is not a plant.
What can actually be claimed
Ten thousand tonnes of annual magnet capacity in 2028 sits against Chinese output somewhere north of 300,000 tonnes. It is enough to cover American defense requirements with margin, and it is nowhere close to enough to cover the automotive, wind and consumer electronics demand that sets the price. Superpower is the wrong word for it by an order of magnitude.
The defensible version of the claim is narrower and still significant: by the end of the decade the United States can plausibly stop being blackmailable on defense-grade permanent magnets. That is a real strategic outcome. It is not mineral dominance, and the people running these companies know the difference even if the podium language does not.
The risk to that narrower outcome is not political. Money is appropriated, contracts are signed, and the bipartisan appetite for cancelling a Pentagon-backed supply chain that Beijing is openly attacking is close to zero. The risk is a supply chain problem one level up from the one everyone is watching. Domestic capacity for the capital equipment itself is the unsolved piece, it has received a fraction of the attention and a smaller fraction of the funding, and it is the variable that decides whether 2028 means anything.
Watch the equipment orders, not the groundbreakings.
