Three developments in September 2026 illuminate a rare earth supply chain under simultaneous pressure from price signals, federal financing, and frontier technology. Ex-China terbium has reached $4,500 per kilogram while EXIM Bank eyes $750 million for Aclara's Louisiana separation hub and USA Rare Earth deploys quantum machine learning to crack the chemistry bottleneck. Together, they reveal how capital and innovation are converging on the one constraint that matters most: heavy rare earth separation capacity outside China.
Introduction
Three numbers define the rare earth market in late September 2026: $4,500, $750 million, and zero. The first is the approximate price per kilogram of terbium oxide outside China, a level that would have seemed fantastical before Beijing imposed export licensing controls in April 2025. The second is the potential ceiling on U.S. Export-Import Bank financing for Aclara Resources' Project Dynamo separation hub in Louisiana, announced in early September. The third represents the number of commercial-scale heavy rare earth separation facilities currently operating anywhere in the Western Hemisphere capable of reliably producing dysprosium, terbium, and yttrium at magnet-grade purity.
These three figures are not coincidental. They are causally linked in a chain that runs from Beijing's Ministry of Commerce through commodity markets, into federal financing offices, and now into quantum computing laboratories. The January 1, 2027 Pentagon ban on Chinese-origin rare earth materials in defense systems is the deadline that makes the chain visible, but the underlying structural deficiency it is exposing has been building for decades.
Understanding how Aclara's Louisiana plant, the USA Rare Earth partnership with Pasqal and Riven Systems, and the deepening ex-China price premium fit together requires stepping back from each individual announcement and reading them as symptoms of the same market failure, now finally attracting responses commensurate with the scale of the problem.
The Price Signal: What $4,500/kg Terbium Is Actually Saying
Price bifurcation in commodity markets is not unusual, but the degree of bifurcation in heavy rare earths has reached a level that belongs in a different category entirely. As of September 2026, terbium metal trades near $1,091 per kilogram on China's domestic Shanghai Metals Market benchmark. Outside China, terbium oxide has reached approximately $4,500 per kilogram. For dysprosium, the gap is similarly stark: domestic Chinese pricing sits around $235 per kilogram, while CIF Europe pricing for dysprosium oxide surged more than 32 percent in August alone to reach approximately $3,500 per kilogram by September.
George Ingall, senior rare earth price analyst at Benchmark Minerals, framed the situation directly: "There is mounting evidence that ex-China heavy rare earth supply chains remain acutely under-supplied relative to demand, a dynamic reflected in pricing across the quarter. Chinese export restrictions continued to constrain availability, with buyers paying significant premiums to secure larger volumes of dysprosium and terbium." Japan, the largest rare earth magnet producer outside China, received just four percent of its normal dysprosium import volumes in the twelve months following the controls. Germany received none.
Yttrium, sometimes treated as a secondary concern relative to dysprosium and terbium, has seen even more dramatic percentage moves. Argus data cited by Reuters puts yttrium oxide prices outside China at 140 times their pre-restriction levels. China controlled approximately 99 percent of global dysprosium oxide and terbium oxide production as recently as 2025, and the single significant non-Chinese refinery, a Vietnamese facility, has been offline for over a year due to a tax dispute. The result is effectively a monopoly that no amount of Western diplomatic posturing has yet dislodged in physical supply terms.
For benchmark NdPr oxide, the picture is more nuanced. Prices opened 2026 near $53 per kilogram, surged above $120 per kilogram by July, and have since corrected to approximately $96 to $97 per kilogram in September. That correction puts NdPr below the $110 per kilogram price floor embedded in the Department of Defense's offtake agreement with MP Materials, meaning the floor is actively in the money. MP disclosed $17.6 million in price-protection income from that arrangement in its most recent reporting period. As I analyzed in September in the context of the Pentagon's broader supply chain maneuvering, the DoD's willingness to write price-floor guarantees reflects an explicit acknowledgment that market prices alone cannot yet sustain the investment levels required for a credible non-China supply chain.
The Capital Response: EXIM's $750M Letter and the Architecture of Federal Support
Against that price backdrop, the U.S. Export-Import Bank's September 4 Letter of Interest for Aclara Resources' Project Dynamo is best understood not as an isolated financing decision but as the latest brick in a federal support architecture that has now reached approximately $3.5 billion in earmarked or finalized commitments during 2026 alone.
Project Dynamo is being developed at the Port of Vinton, Louisiana, roughly 340 kilometers west of New Orleans. Designed to separate Aclara's Super Pure Rare Earth Carbonate into individual rare earth oxides, including NdPr, dysprosium, terbium, samarium, gadolinium, and yttrium, the facility is estimated to cost $277 million to build at its initial commercial scale. The EXIM Letter of Interest covers up to $750 million under the bank's Make More in America initiative, with a repayment tenor of up to 15 years, suggesting the bank is prepared to finance not just the separation plant itself but potential downstream expansions into metals and alloys production. Louisiana has separately committed $46.4 million in state incentives and grants.
What distinguishes Project Dynamo from many announced facilities is its supply chain depth. Aclara's feedstock will not come from a hypothetical future mine. The company's Carina ionic clay deposit in Brazil's Goiás state has completed a prefeasibility study projecting 149 tonnes of dysprosium, 25 tonnes of terbium, and 1,170 tonnes of NdPr annually over an 18-year mine life. The Penco Module in Chile has cleared a permitting process that ran for more than four years. Both sites use Aclara's patented circular mineral harvesting process, which recirculates water and eliminates blasting, crushing, and milling, reducing the environmental footprint that has complicated permitting for conventional rare earth projects.
Critically, the separation technology is not purely theoretical. Aclara inaugurated a rare earth separation pilot plant at Virginia Tech in Blacksburg in March 2026, validating its proprietary solvent extraction process using actual feedstock from its South American deposits. Chairman Eduardo Hochschild noted at the opening that demonstrating separation of Chinese-restricted heavy rare earths inside the United States, using Aclara's own ionic clay supply, was a milestone precisely because it closed the loop that most Western separation proposals leave open: where does the feedstock actually come from?
Basic engineering by Hatch is in its final stages, permitting activities are approaching completion, and Aclara is targeting construction-readiness by end of 2026, with operations potentially beginning in 2028. The EXIM financing, once formalized subject to the bank's standard conditions around engineering documentation, environmental assessments, and offtake terms, would represent one of the largest single federal financing commitments to a rare earth midstream project in American history. CEO Ramón Barúa described it as "a strong validation of our strategy to build an integrated mine-to-alloy rare earth supply chain," and in the context of what that chain is designed to supply, the strategic logic is straightforward: the January 2027 defense ban on Chinese-origin materials requires traceable, non-Chinese separation capacity, and Project Dynamo is among the few projects globally with a realistic path to delivering heavy rare earth separation at commercial scale within a relevant timeframe.
The Technology Bet: Quantum Machine Learning and the Separation Chemistry Bottleneck
Capital and permitting can build facilities, but they cannot by themselves solve the chemistry problem at the core of rare earth processing. Separating a mixture of chemically similar lanthanides into individual high-purity oxides remains one of the most technically demanding and economically punishing steps in the rare earth value chain. Current industrial separation relies on solvent extraction, a process that typically requires hundreds of stages of mixer-settler equipment, enormous volumes of organic solvents, and extractant molecules whose selectivity for specific rare earths has been empirically optimized over decades. China's dominance in separation is not primarily a function of ore reserves but of this accumulated process knowledge, refined at scale across facilities that dwarf anything currently operating in the West.
The September 17 announcement of a strategic partnership among USA Rare Earth, French neutral-atom quantum computing firm Pasqal, and industrial AI specialist Riven Systems represents a fundamentally different approach to closing that gap. Rather than replicating China's empirical trial-and-error accumulation of extractant knowledge, the partnership is attempting to compress the discovery timeline through a three-stage computational and experimental loop.
Riven's self-driving laboratory will conduct thousands of automated experiments generating training data on extractant selectivity across heavy rare earth elements. Pasqal's neutral-atom quantum processing unit will then benchmark quantum machine learning models against classical computing models, both trained on Riven's experimental data, to identify which molecular candidates show the highest predicted selectivity for dysprosium, terbium, yttrium, and other targets. Top candidates will be validated at USA Rare Earth's research facility in Wheat Ridge, Colorado, and integrated into commercial processing flowsheets targeting the company's Round Top deposit in Sierra Blanca, Texas, as well as third-party mixed rare earth carbonates.
The Wheat Ridge facility, commissioned in June 2026, is one of the few Western laboratories currently running parallel separation campaigns on multiple feedstock types simultaneously, including ore, third-party carbonate, and recycled magnet swarf. The facility's 28 engineers and scientists operate around the clock with live SCADA monitoring, and the plant is developing a digital twin in partnership with the Department of Energy's National Energy Technology Laboratory. The current research emphasis on minimizing organic solvent use directly addresses the environmental and cost disadvantages that have historically made Western separation uncompetitive at scale.
Pasqal's quantum computing platform is built on Nobel Prize-winning research in neutral-atom physics, and the company's collaboration with USA Rare Earth represents one of the first deployments of quantum machine learning in a real-world critical minerals application. The strategic logic is compelling: if the quantum ML pipeline can identify extractant molecules with materially higher selectivity than current industrial options, even a modest improvement in separation efficiency could enable smaller, less energy-intensive, and lower-cost processing facilities. That matters enormously in the economics of ex-China rare earth processing, where current facility costs are a primary barrier to competitiveness against Chinese producers operating at scale with decades of optimized process knowledge. USA Rare Earth's chief executive Barbara Humpton has framed the broader company strategy around building the world's leading globally integrated, non-China critical mineral technology platform, and the Pasqal-Riven partnership places frontier technology at the center of that ambition rather than treating it as a secondary consideration.
Convergence: How Price, Capital, and Technology Are Targeting the Same Chokepoint
The surface-level story of September 2026 in rare earths is three separate announcements from three different companies addressing three different aspects of the supply chain problem. The deeper story is that all three converge on a single physical and economic reality: the separation of heavy rare earths from mixed carbonate streams is the chokepoint that everything else flows through, and China's April 2025 licensing regime has priced that chokepoint into global markets with brutal clarity.
Aclara's Project Dynamo targets that chokepoint with infrastructure: a commercial-scale separation facility fed by vertically integrated ionic clay supply from South America, engineered by Hatch, and financed by EXIM. USA Rare Earth and its partners target the same chokepoint with chemistry: a quantum machine learning pipeline designed to discover extractant molecules that could make future separation facilities smaller, cheaper, and more selective than what current industrial practice allows. The price data targets no one; it simply records the consequences of the chokepoint being essentially unaddressed in the Western world for the better part of three decades.
The January 2027 Pentagon deadline functions as a forcing mechanism across all three dimensions. As noted in my earlier analysis of how the November 2026 export control expiration and January 2027 defense ban are compressing the timeline for Western separation capacity, the structural gap between announced ambition and commercial throughput remains the defining vulnerability in the U.S. rare earth position. Project Dynamo's 2028 operations target is at the outer edge of what the defense supply chain needs; the quantum ML partnership's timeline is longer still, aimed at the flowsheets of future facilities rather than the compliance needs of 2027.
That temporal mismatch does not make either project strategically irrelevant. The defense ban creates the immediate compliance crisis, but the permanent competitive challenge of ex-China rare earth processing extends well beyond 2027. The price bifurcation that has pushed ex-China terbium to $4,500 per kilogram is not a temporary disruption; it reflects a structural supply deficit that will persist until Western separation capacity reaches a scale capable of absorbing meaningful volumes of non-Chinese feedstock. Benchmark Minerals' George Ingall described the light rare earth picture as stabilizing but the heavy rare earth situation as remaining acutely undersupplied, and there is no credible near-term scenario in which that changes without both capital-intensive infrastructure investments like Project Dynamo and technology-driven cost reductions like those the quantum ML pipeline is pursuing.
The broader federal financing context reinforces the point. The approximately $3.5 billion in U.S. government support earmarked or finalized in 2026 for rare earths development is historically unprecedented, but it is also a recognition that market prices, even at the extraordinary premiums now prevailing outside China, cannot alone overcome the combination of technical complexity, long lead times, and capital intensity that has deterred Western investment for decades. EXIM's Make More in America initiative, the DoD's price-floor guarantee for MP Materials, the $400 million DoD preferred-stock purchase in MP, and the Mine of the Future grants I covered in September are all instruments of the same policy judgment: that rare earth separation is a strategic infrastructure problem requiring public capital, not merely a market opportunity awaiting private investment.
What Comes Next: Timelines, Tensions, and the Remaining Gaps
The honest accounting of where the Western rare earth supply chain stands in late September 2026 requires acknowledging both the genuine progress and the distances yet to travel. Aclara's pilot plant in Virginia is separating heavy rare earths from its own feedstock. USA Rare Earth's Wheat Ridge facility is running live separation campaigns. The prices that made these projects commercially attractive are demonstrably real and, if China's licensing regime persists, structurally durable. Federal financing at a scale that would have been politically inconceivable five years ago is now flowing.
But no Western heavy rare earth separation facility is yet operating at commercial scale. Project Dynamo's construction-readiness target is end of 2026; first production is 2028 at the earliest. The Carina mine in Brazil targets mid-2028 for operations. USA Rare Earth's Round Top mine, under its Accelerated Mining Plan, targets 40,000 metric tonnes per day of feedstock extraction by 2030. The quantum ML extractant discovery pipeline is at its earliest stages, with validation work ahead before any new chemistry could be incorporated into commercial flowsheets. The gap between announced ambition and physical throughput remains the defining vulnerability.
On the geopolitical side, the tensions are compounding rather than easing. China has added both MP Materials and USA Rare Earth to its export control list, restricting Chinese companies from supplying those firms with dual-use goods and technologies. Defense contractors including Lockheed Martin are overhauling magnet supply chains that require traceability to the mining level, a compliance burden that most multi-tier supplier networks are structurally unprepared to meet by January 1, 2027. The November 10, 2026 expiration of China's suspended export controls adds another near-term inflection point whose outcome will further shape pricing and investment flows.
The most important observation may be that the rare earth supply chain contest is no longer being fought solely in geology or geopolitics. It is now being fought simultaneously in hydrometallurgical laboratories, federal financing offices, quantum computing centers, and commodity markets. Aclara's ionic clay chemistry, USA Rare Earth's quantum ML partnership, and the price signal screaming from ex-China terbium markets are different expressions of the same underlying recognition: the separation bottleneck is real, it is expensive, and it cannot be wished away by policy declarations alone. The question for 2027 and beyond is whether the combination of unprecedented federal capital, frontier technology, and market-driven urgency can close the physical gap before China's structural advantage in heavy rare earth processing becomes permanently embedded in Western industrial supply chains.
Conclusion
September 2026 has delivered three data points that, read together, constitute the most complete picture yet of where the Western rare earth supply chain actually stands: a price bifurcation that has made ex-China terbium roughly four times its Chinese domestic price, a federal financing signal that puts $750 million behind a Louisiana separation hub with genuine feedstock supply, and a technology partnership that places quantum machine learning inside the extractant discovery process for the first time.
None of these developments resolves the fundamental problem. Heavy rare earth separation outside China remains commercially embryonic, the January 2027 defense deadline is closer than any of the major facility timelines, and the chemistry challenge that has kept separation economics firmly in China's favor for decades is not solved by a partnership announcement alone. But the combination of price signal, capital deployment, and technology investment now converging on the same chokepoint represents a qualitative shift from the rhetorical commitments that have characterized Western rare earth strategy for most of the past decade.
For industry professionals and investors, the practical question is sequencing: which pieces of infrastructure come online fast enough to matter for near-term compliance, and which are building the foundation for the decade after. For researchers in separation chemistry, materials science, and quantum computing, the message is that the problem is now commercially urgent enough to attract serious capital and frontier computational tools simultaneously. For policymakers, the lesson of September 2026 is that price signals, financing instruments, and technology investment are all necessary but none is sufficient on its own. The rare earth supply chain outside China will be built in layers, over years, by the convergence of all three, and the window for that convergence to matter strategically is narrower than any of the timelines currently on the table.