Felix Gold Limited (ASX: FXG) told the market on 27 August 2026 that it had built, run and shut down a pilot plant reproducing the flowsheet its Delaware subsidiary Frontier Antimony Refinery Corp. intends to build in the United States, and that ore from the Bundtzen Vein at Treasure Creek in Alaska had come out the other end as cast antimony ingot. The company described the exercise as removing the largest technical unknown ahead of refinery feasibility work, and was equally explicit about what it did not do.
The feed was hand-sorted massive stibnite recovered during bulk sample extraction under Alaska permit APMA F20252839 (Amendment 1). Bags were weighed on receipt and jaw crushed separately without blending, so that traceability held bag by bag, and splits were kept from each for head assay. From there the material went straight to the furnace. There was no grinding, no flotation and no concentrate stage, which Felix Gold attributes to ore quality rather than to any engineering choice.
The four stages that ran
Crushed ore was melted in an externally gas-fired silicon carbide crucible furnace at roughly 1,100 degrees Celsius with no fluxes added, and an air lance above the bath drove oxidation to completion, confirmed by sulphur dioxide generation. The antimony-bearing fume was drawn to a cartridge bag filter and recovered in a baghouse hopper as antimony trioxide, indicatively around 75 per cent antimony by handheld XRF.
Reduction ran in two passes. A scoping stage reduced fume with coal in the gas-fired furnace at 1,100 degrees Celsius under a low-temperature borax and sodium carbonate flux, producing ingots of 127 g and 1,962 g assaying 95.1 per cent in a draft mineralogical report. A confirmation stage used an induction furnace at 1,200 degrees Celsius with briquetted feed in a standard iron oxide and silica slag system, which Felix Gold says fully simulates commercial operation, and produced a 145 g ingot at an indicative 97 per cent antimony. Molten metal was then cast into a mould, and the company noted that briquetted feed under a prepared slag lost materially less fume than the scoping test.
The fifth stage in the design flowsheet, a second refining furnace that would take reduced metal to final specification, was not attempted. Felix Gold states that no metal at 99.65 per cent antimony or any other commercial specification has been produced, that head, fume and metal samples are with an independent commercial laboratory with results pending, and that the XRF grades quoted are indicative field readings not calibrated to certified reference standards. The program was designed and supervised by MetaMets and run at an independent third-party metallurgical facility. Wayne Anderson, an independent consultant and Member of the Australasian Institute of Mining and Metallurgy, signed off the metallurgical information.
The structure the work feeds
Frontier was established in June 2026 as a wholly owned Delaware corporation. No mineral assets were transferred into it; Felix Gold retains Treasure Creek and its gold ground, and Frontier buys antimony-bearing material from the parent under commercial ore supply arrangements. The company says a completed site-selection and permitting study by Worley and SLR has narrowed evaluation to specific priority locations, that preliminary engineering is substantially advanced, and that a seed funding process for Frontier is underway.
On 19 August the US Department of Energy selected a Felix Gold subsidiary for negotiation of an award of up to US$18 million, one of nine projects chosen from a US$162 million program announced on 18 August, with a proposed company cost share of 26.8 per cent, about US$4.8 million, to be delivered in kind as ore. The applicant is Felix Gold Alaska Treasure Creek Inc., not Frontier. Felix Gold has said it is targeting mining operations in June 2027 and a US smelter by end-2027, while labelling both as aspirational rather than production targets.
Analysis: what a batch pilot proves about a continuous plant
The value of this program lies in sequence, not scale. Every commercial antimony route has to move the metal through oxidation, capture and reduction in a specific order, and each handoff creates a way for the process to fail: fume that escapes capture, flux chemistry that carries contaminants forward, feed that will not hold together in a furnace. Running all of it once, on real Treasure Creek ore, converts those from open questions into observed behaviour. Felix Gold says flux behaviour, fume capture, briquetted feed handling and contamination pathways all feed directly into the engineering now under way, which is the concrete use of the result.
The limits are stated in the announcement and are substantial. The plant ran in batch mode on bench and demonstration-scale static equipment, not the continuous, mechanically agitated configuration the design flowsheet specifies, and the company warns results at this scale may not be replicated commercially. No recovery, mass balance, throughput, product specification or production target is reported, and cannot be, because feed grade is still unknown pending laboratory assays. No mineral resource for antimony has been declared at Treasure Creek under the JORC Code, no economic study has been completed, and Felix Gold makes no claim that commercial production will follow.
The absence of a concentrator is the piece worth watching. Skipping grinding and flotation removes capital, unit operations and ramp-up risk, and it is the main reason the company can argue for a shorter build than conventional antimony projects. But it also makes the whole economics dependent on ore that stays clean enough to charge directly, hand-sorted and coarse-crushed, at commercial tonnages. The pilot was fed on hand-sorted massive stibnite, which is not the material a continuous mining operation delivers. Whether the same route holds on continuously mined material is a question the bulk sample program, not this pilot, will answer.
The independent numbers explain the urgency. USGS reported no marketable antimony mined in the United States in 2024, net import reliance of 85 per cent of apparent consumption, and world mine production of about 100,000 t of which China accounted for 60,000 t, Tajikistan 17,000 t and Russia 13,000 t. After China required export licences in August 2024, USGS records the metal price nearly doubling from $8.91 per pound in July to $17.50 per pound in November, then a December 2024 ban on antimony exports to the United States. A refinery with a domestic anchor feed addresses that supply position, and the federal funding program described below was announced in the same context.
What would confirm the result
Three items sit ahead. Independent laboratory assays on head, fume and metal, which will replace the XRF readings and allow a recovery figure for the first time. The second refining stage, which is what would demonstrate metal at specification rather than the indicative purities reported here. And a final US site, without which Felix Gold says no capital or operating cost estimate can be released, since site-specific infrastructure, permitting and logistics inputs all remain outstanding. The balance of the crushed sample is planned to follow the same route with a single larger ingot cast.
