USA Rare Earth: Dysprosium Prices Up 90% as CEO Barbara Humpton Hands Off Mine-to-Magnet Empire With $1.5 Billion Cash Cushion
Q2 2026 earnings call, August 10, 2026
USA Rare Earth used its second-quarter call to make a blunt case that a structural, two-tier pricing regime has taken hold in rare earths, with Western buyers now paying multiples of Chinese prices for the heavy elements that go into magnets, semiconductors and defense systems. Outgoing CEO Barbara Humpton told investors that Western dysprosium oxide prices are up more than 90% in 2026 alone, "reaching nearly $2,000 per kilogram in August, over 9 times the price of the product in China," a figure she attributed to Benchmark Mineral Intelligence. Yttrium oxide, a metal the company only recently began producing commercially, has risen 60% since March and now trades at more than 200 times the Chinese price. The framing is central to the investment case: this is no longer a story about who has the lowest-cost tonnage, but about who can guarantee availability at any price.
That scarcity is a double-edged sword. CFO Rob Steele acknowledged gross margins in the metal-and-alloy business were squeezed by higher raw material input costs tied to the same heavy rare earth shortages hitting the entire industry, forcing the company to source alternative feedstock ahead of its Serra Verde and Carester supply coming online. But management is leaning into the tightness rather than hedging against it. "Our position in the supply-constrained market conditions provides the opportunity to establish appropriate non-China pricing," Steele said, adding that the company has already gone out with higher prices on its products and expects that to show up in revenue "in the upcoming quarters."
Serra Verde acquisition nears the finish line
The most consequential near-term catalyst is the pending acquisition of Serra Verde, which would give USA Rare Earth the only scaled operating source of both light and heavy magnetic rare earths outside Asia. Management confirmed there are no remaining regulatory hurdles; the sole remaining condition is a shareholder vote on August 28, with closing expected shortly after. Analyst George Gianarikas pressed on the operational risks behind Serra Verde's target of roughly $600 million in annualized EBITDA by the end of 2027. Steele pointed instead to a more specific production target: 6,400 metric tons of run-rate TREO capacity by the end of 2027, with more detail promised once the deal formally closes. The strategic logic is straightforward — Serra Verde is a heavy rare earth mine that, combined with Round Top's heavy-only resource, is meant to give the company a genuinely non-Chinese heavy rare earth backbone, something almost no other Western company can claim.
Commercial pipeline is real but still mostly unpriced
The company disclosed it is in active commercial dialogue with more than 100 potential customers, with over 20 already in qualification and MOUs and LOIs covering 2,500 metric tons of annual demand signed with large multinationals across aerospace and defense, industrial automation, industrial motors and automotive. Some of those have already converted into production purchase orders, though management declined to quantify how much. Steele was careful to note that qualification timelines vary widely by customer and even by product line within the same customer, but reiterated that with production POs already in hand, the company expects its first magnet sales by year-end. Investors should treat that guidance as directional rather than firm, given the acknowledged variability in qualification cycles and the fact that none of the conversion volume has been quantified.
Capacity build-out: Stillwater first, Blacksburg later
Stillwater, the company's primary magnet and metals facility, is targeted to reach 600 metric tons of run-rate capacity by year-end, doubling to 1,200 metric tons in the first quarter of 2027, en route to an eventual 3,600 metric tons of magnet-making and 5,000 metric tons of metal-making capacity. The newly selected Blacksburg, South Carolina site, where the company has broken ground, will not be operational until early 2028, with its building shell not even complete until the end of 2027 — a reminder that the platform's most ambitious targets, including 10,000 metric tons of combined U.S. metal, alloy and magnet capacity by 2029, remain several years out and dependent on sequential execution across two greenfield sites plus an international acquisition.
Recycling economics could meaningfully lower the cost base
One underappreciated detail: Steele disclosed that magnet manufacturing "swarf" — the scrap generated during finishing — typically represents 20% to 30% of finished magnet output, meaning a 10,000-metric-ton production target could generate a couple thousand metric tons of recyclable swarf annually. The company already produced its first commercial-grade dysprosium and NdPr oxide samples from its own recycled swarf in July, a technically demanding process it says few Western producers can execute outside Asia. If scaled, this recycling loop could supply as much as 20% to 30% of total feedstock needs, reducing dependence on newly mined material and potentially improving unit economics as volumes ramp.
Balance sheet and cash burn
The company ended the quarter with roughly $1.5 billion in cash, giving it flexibility to fund both organic buildout and further M&A. Revenue remains minimal at approximately $6 million, entirely from third-party metal and alloy sales at LCM, against operating expenses of about $45 million, inflated by M&A, legal and consulting costs tied to the Serra Verde, Carester and Blacksburg transactions. The company reported a net loss attributable to common stockholders of $10.3 million, or $0.05 per share, though this included a $22.4 million noncash fair value gain on warrant and earn-out liabilities; excluding that item, the adjusted net loss was $33.5 million, or $0.15 per share. Capital expenditures were $66 million for the quarter. Notably, the Department of Commerce financing signed this quarter is structured as milestone-based capex reimbursement — the company is paid only after hitting specific targets — which management frames as taxpayer protection but which also means the cash benefit is back-loaded and contingent on execution.
Leadership transition adds execution risk to watch
Humpton, who has led the company for less than a year, will hand the CEO role to Thras Moraitis on October 1, calling this her final quarterly call in the role. She was candid that the talent buildout has been substantial — from "tens of employees" when she joined last fall to several hundred now, bolstered by the Serra Verde, Carester and LCM teams joining the platform — but the timing of a CEO transition just as Serra Verde closes and Stillwater ramps toward first magnet sales is a governance variable investors will want to monitor closely, even with Humpton describing her confidence in the handoff as high.