Energy Fuels’ Heavy-Rare-Earth Build Makes Donald Feedstock the Next Bottleneck

Energy Fuels has moved its rare-earth strategy from acquisition slides into construction. The company has started building commercial heavy-rare-earth separation circuits at the White Mesa Mill in Utah, with dysprosium and terbium oxide production targeted by the end of 2027 and samarium, europium and gadolinium circuits scheduled a year later.

That changes the central execution question for Energy Fuels. Its proposed $1.9 billion acquisition of VAC gave the company downstream magnet capacity and long-standing customers. Construction at White Mesa now begins to close the midstream gap. The constraint increasingly shifts upstream: can Energy Fuels secure enough monazite, on time and at the right composition, to keep those separation circuits and downstream plants supplied?

The answer is increasingly tied to the Donald project in Victoria, Australia. Energy Fuels said with its second-quarter results that Donald could reach a final investment decision as early as the third quarter of 2026. The company owned 12.7% of the joint venture at June 30 after contributing A$48.83 million, retains the option to earn up to 49%, and holds rights to 100% of the project’s monazite offtake.

Donald is expected to produce roughly 8,500 to 9,500 tonnes of monazite concentrate a year from 2028. Astron, Energy Fuels’ joint-venture partner, describes phase-one rare-earth concentrate output at about 9,000 tonnes annually, rising to 13,000 tonnes in a second phase. The modeled mine life is 39 years. Critically for White Mesa, the concentrate is expected to contain meaningful dysprosium, terbium and samarium rather than being merely another light-rare-earth feed.

That matters because heavy rare earths remain the technically and commercially harder part of the magnet chain. White Mesa already produces neodymium-praseodymium oxide and Energy Fuels has demonstrated pilot-scale terbium production, but sustained commercial separation of dysprosium and terbium requires additional solvent-extraction capacity, feed consistency and impurity control. The current build is therefore more consequential than another pilot campaign: it is designed to turn White Mesa into a plant capable of routinely separating both light and heavy rare-earth oxides.

Energy Fuels expects Donald’s first-phase material, together with third-party feed, to fully utilise its current NdPr capacity and planned Tb and Dy circuits. The company says that oxide output could provide about 70% of the feed required for Australian Strategic Materials’ existing and planned metals-and-alloys capacity in South Korea. Those alloys, in turn, could supply more than the current 2,000-tonne annual permanent-magnet capacity at VAC’s Sumter, South Carolina plant if the pending acquisitions close.

The numbers show why feedstock timing matters. Energy Fuels’ larger Phase II White Mesa plan targets up to 6,000 tonnes a year of NdPr oxide, about 288 tonnes of dysprosium oxide and 80 tonnes of terbium oxide. That scale is substantially larger than the current system and is intended to support an eventual expansion of VAC’s Sumter magnet plant toward 12,000 tonnes a year. A delay at Donald or another upstream project would leave an expensive separation-and-magnet platform more dependent on third-party monazite and mixed rare-earth carbonate purchases.

Financing is less constrained than it was a year ago, but it is not irrelevant. Energy Fuels ended June with about $996 million of working capital, including $58 million of cash and $878 million of marketable securities. It also has a conditional commitment for up to $725 million of 20-year senior secured debt from the U.S. Office of Strategic Capital to support White Mesa expansion and a planned U.S. metals-and-alloys facility. The commitment still requires due diligence, final documentation and closing conditions.

The buildout is already affecting the income statement. Energy Fuels posted a $33.6 million second-quarter net loss, compared with $21.8 million a year earlier, as transaction expenses and operating investment rose. Uranium remains the cash-generating foundation: the company sold 310,000 pounds of U3O8 in the quarter at an average realised price of $80.48 per pound and reported a weighted average mining, transport and milling cost of about $23 per recovered pound for Pinyon Plain ore processed in its latest conventional mill run.

That uranium margin gives Energy Fuels an advantage over rare-earth developers that must finance their entire processing chain before earning meaningful revenue. But it does not remove integration risk. The company is simultaneously trying to close the ASM and VAC acquisitions, commission new heavy-rare-earth separation, advance Donald, and ultimately connect oxides to metals, alloys and magnets across the United States, Australia, South Korea and Europe.

Washington’s latest push to fund domestic critical-mineral projects improves the policy backdrop, and Energy Fuels already has one of the more tangible U.S. processing platforms. The next proof point, however, is industrial rather than political. Donald needs to reach FID and stay on schedule, while White Mesa must demonstrate that its new circuits can produce specification-grade Dy and Tb oxide continuously at commercial scale.

If both happen on time, Energy Fuels will have something still rare outside China: secured heavy-rare-earth feed, commercial separation, metal and alloy conversion, and qualified magnet manufacturing. If Donald slips, the company may discover that building downstream capacity was easier than feeding it.