In mid-2026, spot FOB China for 99% min cerium oxide sits near US$5,100 per ton while 99% min praseodymium oxide (Pr6O11) lists at US$35,500 per ton on Chinese export platforms, a roughly 7:1 spread that captures how light rare earths are still treated as a high-volume industrial mineral and the medium-heavy oxides as a specialty feedstock [S2].
Adamas Intelligence opened registration on 2024-07-15 for its Rare Earth Mines, Magnets & Motors 2026 conference (Toronto, 2026-09-28 to 2026-09-30), alongside a new Rare Earth Market Monthly price service covering oxides, metals, alloys, and NdFeB alloys, indicating that the analytical market now treats REE pricing as a continuously tracked benchmark rather than a quarterly report [S1]. On the technology side, Rare Earth Technologies, Inc. (RETi) is promoting a non-HPLC, single-pass column process with claimed purity up to 99.999% REE, using mineral acid and water only as solvent, with a three-year-plus maintenance cycle and equipment supplied by YMC America [S3].
Where the value sits: oxide pricing bands from a Chinese export desk
Zhejiang Jinda Rare-Earth Element Co., Ltd. lists 99% min rare earth cerium oxide (CeO2) powder at FOB US$5,100/ton with a 1-ton minimum order, and high-quality praseodymium oxide Pr6O11 at FOB US$35,500/ton, also on a 1-ton minimum order, on a 2026 product page with five total SKUs [S2]. The 1-ton minimum on both grades is a direct signal that midstream Chinese merchants still transact in full-pallet-to-container volumes rather than lab sample sizes, which matters for downstream specifiers sizing flow-meter and pressure-transmitter instrumentation on hydrometallurgy trains.
Two structural facts can be read from the price spread. Second, the praseodymium-neodymium pair and other magnet-relevant oxides carry a 5-10x multiplier, so any process change that loses Pr/Nd yield to tails has a disproportionate cost impact per ton of mixed concentrate fed.
The separation gap RETi is targeting
RETi, formed in 2021, positions itself as a concentrate processor and refiner rather than a miner, taking mine-derived or recycled concentrates and outputting separated oxides or carbonates at claimed purity up to 99.999% in a single column pass, with no toxic waste, using mineral acid and water only as solvent [S3]. The company explicitly states it is not currently licensing its process, meaning any Western OEM that wants this chemistry in 2026 must source separated oxides from RETi itself or build a parallel solvent-extraction train, which is the real procurement decision behind the announcement.
The headline 99.999% figure matters for two engineering reasons: NdFeB sintered magnet makers typically need Nd/Pr oxide assays at 99.5% or higher to hit grain-boundary control, and optical-grade cerium polishing slurries reject iron and thorium at the ppm level, so the spec wall is set by the magnet line, not the separator. RETi's three-year-plus maintenance cycle claim, if delivered, would offset the typical 3-6 month solvent-extraction cleanout cycle that drives working-capital costs at conventional Chinese separation hubs [S3].
Policy and demand drivers pulling on the 2026 market
Adamas Intelligence is publishing a free report titled Humanoid Robot Revolution: NdFeB Supply Chain Implications and Strategies, framing the magnet-relevant REE outlook around the metal intensity of 10 billion humanoid robots, and a separate report on the contrasting industrial policies of the US, EU, and China, plus a viability study of the EU's Critical Raw Materials Act [S1]. Three concrete signals sit in those titles: humanoid electrification is being modelled as a NdFeB demand vector, the US/EU/China policy split is now an explicit analytical category, and the CRMA is being stress-tested for deliverability, not aspirational targets.
For a specifier, this converts into a sourcing rule: if your bill of materials contains NdFeB magnets, EV traction motors, or servo motors, treat 2026 as a dual-track year, with Chinese midstream supply on price-competitive volume and ex-China supply on policy compliance, and price both tracks. Adjacent industrial specifiers are reading similar signals in 2026 guides for magnetic flow meter selection, where REE content in flow instrumentation is small but real, and in pneumatic actuator sourcing, where rare-earth-bearing coil and valve-trim alloys are part of the certification gate.
Process selection: SX vs ion-exchange vs single-pass column
Three separation routes are competing for the same mixed-REE concentrate in 2026. RETi's single-pass column route claims 99.999% REE purity, no toxic waste, mineral-acid-and-water-only solvent, and a 3-plus-year maintenance cycle, with YMC America as the equipment supplier [S3]. Classical ion-exchange / HPLC, which RETi explicitly disclaims, achieves very high purity at the cost of throughput and resin lifetime, and is reserved for analytical and small-lamp-phosphor niches.
The comparison sharpens on four decision criteria a buyer can act on. Cost per kg of separated NdPr oxide: SX is lowest at scale, RETi is competitive only where waste-handling capex is priced in, and HPLC is the most expensive by 1-2 orders of magnitude. Environmental footprint: RETi's mineral-acid-and-water-only claim and "no toxic waste" line are the differentiator, while SX carries organic-phase disposal liability. Purity ceiling: RETi 99.999% beats SX's typical 99.5-99.9% product-grade output and matches HPLC [S3]. Throughput per train: SX remains dominant, RETi is unproven at the 10,000-ton scale, and HPLC is laboratory-bound.
Where this hits industrial specifiers directly
Rare earths enter industrial bills of materials in three non-obvious places beyond magnets. Cerium oxide is a polishing slurry for precision optics and semiconductor wafers, where 99% min assay and iron content below spec are the buy gate, and the US$5,100/ton FOB price on Chinese export desks sets the cost ceiling for a competitive alternative [S2]. Lanthanum and cerium are glass-polishing and catalytic-cracking additives, where industrial valve and pressure sensor trim specifiers will see REE-bearing alloys in severe-service trim, and where dual-source qualification is the procurement hedge for 2026.
Praseodymium-neodymium oxide at US$35,500/ton is the line item that justifies any NdFeB magnet audit, and the same oxide price feeds back into the spec pages of EV traction motors, wind turbine generators, and the servo motors on AGV fleets covered in 2026 specifier guides for AGV robot selection. Buyers should price the NdPr oxide line, then run a magnet-content audit on each motor SKU, and only then decide whether RETi-style non-Chinese separation or conventional Chinese SX is the lower-risk 2026 procurement answer [S1][S3].
Selection criteria: who the 2026 REE market is for, and who it is not
This market is for NdFeB magnet buyers who can hold 6-12 months of NdPr oxide inventory, for optical polishers willing to dual-source cerium slurry from a non-Chinese refiner, and for defence and aerospace primes whose magnet bill of materials is a CRMA / US Defence Production Act compliance question. It is not for buyers of single-kilogram research-grade REE, where HPLC chemistry still wins on flexibility, and it is not for downstream OEMs whose magnet content is below roughly 50 g per unit, because the working-capital cost of dual-sourcing outweighs the geopolitical insurance. [S1]
Two verifiable next nodes worth tracking. The Adamas Intelligence Rare Earth Mines, Magnets & Motors 2026 conference on 2026-09-28 to 2026-09-30 in Toronto is where the 2027 baseline forecasts will be published, and the Adamas Rare Earth Market Monthly is the first place oxide price moves between Chinese export desks and Western concentrate processors will be visible as a continuous series [S1]. RETi's first commercial-tonnage delivery outside China, if it lands before mid-2027, is the second node, and is the cleanest signal of whether single-pass column chemistry can break the Chinese SX cost curve at industrial scale [S3].