USA Rare Earth: A Record Dysprosium Basket, a Halved Stock, and a Round Top That Slipped a Year
USA Rare Earth reported third-quarter results on 2026-11-09 into a tape that had already made up its mind. The shares are down 55.5% over the last 90 trading days, having peaked at $20.00 on 2026-08-14 and fallen 60.4% from that high; the full history shows a 133% run, a 54% four-day spike, then a 44% and a 45% collapse — a stock that trades on narrative velocity rather than cash flow. Market cap now sits near $1.05 billion, a fraction of the ~$3.3 billion figures floating around three weeks earlier. So the question for new CEO Thras Moraitis, hosting his first call, was less "what did you do" than "why should we believe the next milestone when the last one moved?"
The basket is spectacular. The revenue is $3 million.
The single best fact in this report is buried in the CFO's fourth-quarter outlook: the Serra Verde basket floor averages about $43 per kilogram, but the products actually being sold are realizing roughly $97 per kilogram quarter-to-date, driven entirely by record ex-China prices for dysprosium and terbium. On an NdPr-equivalent basis the company put that closer to $340/kg. Moraitis pressed the point that “this structural imbalance has driven dysprosium and terbium prices to record levels, a development that, in our view, is not yet fully appreciated by the market” — Thrasyvoulos Moraitis, Chief Executive Officer · 2026-11-09.
So the $97 is basically what we receive. That's the net. And that's after the subtraction of an imputed separation cost and also the share that the SPV gets above the floor price... If you look at it on an NdPr equivalent basis, then we are much closer to $340 per kilo.
And yet quarterly revenue was $3 million, down quarter-on-quarter, because the company could not get the inputs to make things. Dysprosium oxide is described in almost siege terms: “It's literally a hand-to-mouth situation on dysprosium oxide. And so we're picking up small tonnes here and there until, of course, we are able to separate the Serra Verde product in our own facilities.” — Thrasyvoulos Moraitis, Chief Executive Officer · 2026-11-09 Samarium metal is the other bottleneck — the oxide is available, the furnaces are not, and two furnaces on order lift capacity ~30% only around mid-2027.
That is the central incoherence of the quarter and, arguably, of the whole equity story: the basket price is extraordinary precisely because the supply chain is broken, and USA Rare Earth is on both sides of that break. It owns the scarce heavy rare earth feedstock and it also owns the downstream converter that cannot feed itself. Meanwhile the celebrated "first commercial magnet sale" is, on the CFO's own framing, a proof point rather than a business: “we're currently in prototype sales, which means we are proving the fact that we can sell magnets to our customers, but prototype cells are low volumes.” — William Steele, Chief Financial Officer · 2026-11-09 The prior quarter's top keyword was purchase order; this quarter it has fallen out of the top thirty entirely, replaced by magnet qualification — a slower, more honest word.
Round Top's clock moved, and nobody called it a delay
In August, the CFO was unambiguous about the flagship Texas asset: “We still are on track to complete our DFS by year-end and publish our S-K 1300 going into early next year.” — William Steele, Chief Financial Officer · 2026-08-10 Three months later, the S-K 1300 for Round Top is scheduled for the third quarter of 2027, with FID in the second half of next year. The company frames this as rigour — extra drilling, 1,200 of 5,500 assays returned, 72% heavy distribution, WSP as qualified person, Fluor on flow sheet, and a desire to reach 50-60% detailed engineering before committing capital. Moraitis's reasoning is genuinely sound engineering practice.
But it is also a one-year slip on the milestone investors were anchored to, disclosed in the same breath as a new list of key operational targets — eight dated checkpoints across the three new business units, up from essentially none. A new CEO substituting a precise calendar for a vague promise is healthy governance and simultaneously an admission that the old calendar was optimistic. Note too what he chose to pre-empt: “Contrary to speculation from some quarters, production has ramped up meaningfully with volumes, recoveries and product quality all improving at pace.” — Thrasyvoulos Moraitis, Chief Executive Officer · 2026-11-09 You do not rebut a rumour you have not heard.
The discipline is real — and it is a brake
The capital story is where the call most diverges from the bull case. Cash stands at $1.2 billion, against $148 million of operating cash used and $106 million of capex plus equipment deposits in the quarter. That burn is visible in the filings: free cash flow ex-SBC was -$133M in the June quarter, against $70M of capital expenditure and $33M of effective SG&A. The balance sheet is still clean — liabilities at 15.1% of assets — but the equity, not the debt, is what is funding this.
Which makes the Department of Commerce answer the most revealing moment on the call. Asked about the award, Steele noted funds are appropriated, the topic is bipartisan, and then: “we are not spending CapEx as quickly as the equity raise milestone assumed, and we have not submitted a reimbursement request to the DOC.” — William Steele, Chief Financial Officer · 2026-11-09 Read plainly, the company has deliberately decelerated below the pace its own government agreement presumes, is choosing not to draw money it could draw, and is reserving the right to renegotiate the equity milestone. It is prudent. It is also the opposite of the land-grab posture the sector's valuation has been built on.
What has not changed is the strategic logic, and here the company is not alone. MP Materials reported on 2026-11-05 leaning on exactly the same vocabulary — rare earth separation, the heavier earth business, margin opportunity, strategic customers. Two competitors converging on heavies is confirmation that the scarcity is structural rather than rhetorical, and it is the strongest argument for the value chain being assembled across Brazil, the U.K., France and the United States. A year ago management told investors “we feel very good that the current capacity will -- can supply us over the next year to 18 months” — William Steele, Chief Financial Officer · 2025-11-08 on heavies; today the same company describes a hand-to-mouth scramble. That is the distance between a thesis and an operating reality.
The verdict
USA Rare Earth is now a company with the best heavy rare earth price environment in its history, a first sale of NdFeB magnets, a producing Brazilian mine, and $1.2 billion of cash — and its stock has halved in six weeks because the milestones that justify the multiple keep sliding right while the revenue line stays at $3 million. The industry outside China is genuinely being built, and Moraitis's instinct to slow down and engineer properly before FID is the right one for a 2030 outcome. It is the wrong one for a stock priced on 2027. Until the separation capacity exists — his own words put the U.S. plant at early 2028 — the company will keep selling a basket it cannot fully process, at prices it cannot fully capture. Spectacular optionality, deferred again.