Thirty million tons of red mud sits in the bayous of Gramercy, Louisiana. It is the toxic aftermath of alumina refining, the industrial waste that aluminium smelters have been stacking along the Mississippi River for decades. On June 2, the Department of Energy announced it would spend $67 million to turn that waste into rare earth elements. Colorado School of Mines and ElementUSA, a Cedar Park–based minerals processing company, will design, build, and operate a facility extracting dysprosium, terbium, yttrium, gadolinium, neodymium, praseodymium, samarium, and lanthanum from the refinery tailings. The target is 150 to 1,000 metric tons per year of finished rare earth oxides and metals.
This is not a small pilot. The Gramercy facility is one of two projects selected under the DOE's Rare Earth Demonstration Facility program, which carries $134 million in total funding designed to fund advanced extraction and processing plants that can supply U.S. demand from waste streams and non-traditional sources. The other awardee is Phoenix Tailings, which was awarded $66 million and announced its award publicly on June 4, 2026. The scale of federal capital moving here is unmistakable: $67 million in committed funding to build infrastructure, not to subsidize mining permits or study feasibility. ElementUSA already operates a Critical Resource Accelerator lab-to-pilot facility in Cedar Park for process verification and scaling; Colorado School of Mines brings research capacity and supply-chain credibility. This partnership is designed to move from design to commissioning in a reasonable timeline, not to generate academic papers.
The feedstock change is the economic hinge. For thirty years, U.S. rare earth strategy has meant buying refined oxides from China or Australia, or extracting REEs from primary ore at huge cost. Alumina tailings flip the equation. Red mud contains rare earth oxides at concentrations that make extraction viable if processing costs can be pushed down. The Gramercy site is not choosing to locate there; the waste is already there, already stacked, already a remediation liability. ElementUSA is not paying for ore. It is converting a disposal cost into a raw material. If the plant hits the 1,000 MT/year target, U.S. rare earth supply moves from 0 percent domestic rare earth processing today to domestic sourcing of a critical material class, concentrated in the heavy rare earths (dysprosium, terbium) that are the binding constraint on domestic magnet supply.
Heavy rare earths are the real measure here, not the total tonnage. Dysprosium and terbium are the most expensive and the most scarce. They are irreplaceable in permanent magnets used in wind turbines, military radar systems, and advanced electrical machines. China controls 70 percent of global rare earth mining and 85 percent of rare earth processing; the U.S. has zero domestic processing capacity. A Gramercy facility running at even 200 MT/year of mixed oxides would add 10–20 MT/year of heavy rare earths to U.S. supply, which is not transformative on a global scale but is transformative for U.S. dependency. The plant's viability depends on whether ElementUSA can push processing cost per kilogram below what it costs to import Chinese-refined oxides. If the waste-stream feedstock model works, it undercuts primary mining economics and creates an arbitrage that makes domestic processing cheaper than extraction from ore.
The timeline and scaling rate are now the open questions. DOE awards of this size typically fund design and permitting in year one, construction in years two and three, and commissioning by year four or five. Gramercy has existing industrial infrastructure, power, and water access; it is not a greenfield site. But rare earth processing involves chemical precipitation, solvent extraction, and metal reduction, each step with efficiency curves. The proof point is whether the plant demonstrates that 200+ MT/year is achievable and whether unit processing cost falls into the range where domestic supply undercuts imports on economics alone, not subsidies. ElementUSA will need to prove both. Watch the facility's commissioning date (likely 2028–2029), first production announcement, and whether capacity utilization reaches 500+ MT/year within two years of start-up. If scaling stalls below 200 MT/year or processing costs remain above import parity, this becomes a subsidy-dependent model and the strategic case for domestic rare earth processing weakens. If it hits 500+ MT/year and runs at imported parity cost, the economics of domestic critical minerals just shifted.