When China's MOFCOM placed seven medium and heavy rare earths under export licensing on April 4, 2025, yttrium was one item on a list of seven. Sixteen months later it is clear that the same regulation did not land the same way on all seven. Yttrium is the only one of the group consumed as a bulk ceramic additive at commodity volume — scandia does the same percent-level job in a stabilized zirconia at a fraction of the tonnage — which is why it is the item that lands on yttria-stabilized zirconia (YSZ) rather than on magnets or imaging agents. The point up front: the control bites hardest at the oxide tier, blurs at the sintered-product tier, and delivers its cost at the coating-and-service tier — and the reason sixteen months of extraordinary prices produced so little substitution is that the constraint was never money. It was requalification time. This article is a companion to our scandium export-control analysis, which covers the other side of the same announcement; the Korean customs, K-REACH, and contract mechanics are treated separately in our Korea import guide for YSZ, and market size and grade structure in the 2026 YSZ market outlook.
The April 2025 Package, and Why Yttrium Is the Odd One Out
Announcement No. 18 controlled samarium, gadolinium, terbium, dysprosium, lutetium, scandium, and yttrium together, across metal, alloy, oxide, compound, and mixture forms. Treating them as a single bloc is the first analytical mistake, because they differ by two to three orders of magnitude in both volume and dose.
Six of the seven are consumed in gram quantities inside a high-value assembly. For a magnet maker, a controlled input is a cost pass-through problem: a several-fold increase in an additive that is a small fraction of a motor's bill of materials is painful but absorbable. Yttrium is the exception. It is a structural and functional additive that makes up 3 to 8 mol% of a ceramic sold by the kilogram, in a market whose economics are measured in dollars per kilogram. When the additive reprices by two orders of magnitude, it stops being an additive line and becomes the product's cost.
Scandium is the near-analogue in chemistry — percent-level loading in a stabilized zirconia — but its market is roughly 40 tonnes a year against yttrium's tens of thousands, so the disruption there was concentrated in a specialist niche. Yttrium's disruption was commodity-scale, which is a different problem entirely.
The practical asymmetry also showed up in flows. By May 2026, according to trade-flow tracking, dysprosium shipments were running within the range of pre-control levels with most volume going to South Korea, while yttrium exports remained well below pre-control levels and shipments to Japan — historically a major destination — were near zero. Controlled together does not mean constrained equally. The specific classification language that pulls Y-Zr systems into scope, including the explicit naming of yttrium-zirconium alloy targets under 1C908(a) and the absence of any concentration floor, is set out in our Korea import guide and is not repeated here.
Tier by Tier: Where the Control Bites and Where It Blurs
The useful way to read this control is not "is YSZ controlled" but "at which tier does the regulation actually change behaviour."
At the oxide tier the effect was unambiguous. US yttrium imports fell from roughly 333 tonnes over the eight months preceding the controls to about 17 tonnes in the post-control period reported. We deliberately do not restate that as a percentage: we have not been able to confirm that the two windows are the same length, and comparing an eight-month figure against a shorter span would overstate the fall. Prices for yttrium oxide outside China rose about 140-fold from a pre-control base under $8/kg. Shipments of yttrium, dysprosium, and terbium combined were still down roughly 50% versus the twelve months before April 2025.
At the powder tier the control produced a mirror-image squeeze. Chinese powder makers hold a controlled output and need a licence to ship it; non-Chinese powder makers hold a controlled input and need someone else's licence to receive it. Both directions run through the same MOFCOM queue.
At the sintered-product tier, sixteen months produced no published guidance resolving whether a fully sintered YSZ article is "a mixture." Whatever the reason for that, the planning consequence is the same: sixteen months without clarification is long enough that a buyer should treat the ambiguity as durable and schedule around its persistence rather than around a ruling that may not arrive.
The cost landed at the last tier — the coating shops, maintenance operations, and stack builders that the control does not mention at all. Trade reporting through the period described rationing that prioritised large OEM contracts over maintenance shops, at least two aerospace coating manufacturers temporarily idling YSZ lines, and one supplier exhausting its yttrium oxide inventory and halting sales. This tier has the shortest contracts, the least inventory capacity, no pass-through leverage against an OEM, and no authority to change a material specification it did not write. It is the tier furthest from the regulation and the one that felt it most.
What Sixteen Months of Licensing Actually Produced
"Controlled" and "unavailable" are different states, and the gap between them is administrative rather than physical.
There is no statutory approval deadline. Provincial commerce bureaus informally indicated a 45 to 60 day review, while observed processing ran from about 45 days to several months, with applications requiring an end-user certificate, technical description, importer details, and value-chain documentation. Then, in December 2025, MOFCOM confirmed it had approved the first general licences — authorising repeated shipments to pre-approved buyers over a period rather than one application per shipment. Neither the number issued, the validity period, nor the qualifying criteria were published.
The result is a bifurcated market. A buyer with committed volume, a documented civilian end use, and an accumulated compliance record sits behind a general licence and something close to a normal ordering cadence. A buyer placing small spot orders is still filing per-shipment applications with no schedule certainty. This is the same allocation pattern that appeared early in the scandium episode, now institutionalised as policy rather than emerging as informal supplier behaviour.
The measurable output is the dual price, and it needs to be read carefully because the two published benchmarks are not the same product.
Metal and oxide are different forms on different bases, so the ratio between those two figures is not a like-for-like premium and should not be quoted as one. What the pair does establish is direction and order of magnitude: the inside-China and outside-China markets have stopped tracking each other. The domestic metal benchmark eased 9.4% between July and early August 2026 while ex-China oxide stayed at multiples of its pre-control level. Some gap between the two would be expected on fundamentals alone — different form, tax treatment, and freight. What is harder to explain that way is the direction: softening inside the border while the outside stays elevated is the signature of a constraint on getting material across it, not of a change in underlying demand. The full price trajectory sits in our market outlook.
One gap worth naming: no approval or refusal rate specific to yttrium has been published, by MOFCOM or by any source we could verify. Anyone quoting a percentage approval figure for yttrium licences is estimating.
Substitution and Requalification: The Cost Is Time
Buyers tried four things. None of them failed on chemistry.
Grade downshift. Reducing yttria loading, or using a lower-yttria grade, works where the property being purchased is mechanical. It does not work where the yttria content is the function: in 8YSZ the yttria is what holds the cubic phase and delivers ionic conductivity and thermal stability. The strength-grade versus functional-grade distinction is laid out in our market outlook, and it is the line beyond which downshifting stops being a sourcing decision and becomes a product change.
Non-Chinese yttria. Lynas is expanding heavy rare earth output through Mount Weld and its Malaysian processing, MP Materials is developing yttrium at Mountain Pass, and ReElement Technologies has stated a target of 200 tonnes per year of yttrium oxide. Against a 15,000–20,000 tonne global scale, and with China holding roughly 85–90% of capacity, these are directionally important and near-term immaterial; published assessments put meaningful impact in the late 2020s.
Recycling and reclaim. The chemistry is demonstrated, though the routes are grade-specific. Hydrothermal treatment at 200–240 °C spontaneously disintegrates sintered 3Y-TZP into powder — it works by driving the tetragonal-to-monoclinic transformation, so it is a route into tetragonal dental and structural scrap, not into the fully stabilized cubic 8YSZ that carries the SOFC and TBC demand. What does cross grades is chemical recovery: published leaching work reports Zr(IV) and Y(III) extraction efficiencies around 98% and 95%, and recovery of YSZ from end-of-life solid oxide cells including nickel removal has been carried out at laboratory scale. Spent dental zirconia is roughly 92% zirconium and 6% yttrium by oxide content — a richer yttrium source than most ore, though what that route yields is Y₂O₃ feedstock, not a finished cubic electrolyte powder. What does not exist is a merchant-scale reclaim stream, and reclaimed powder faces the same requalification gate as any new source.
Stabiliser change. Ceria, magnesia, calcia, and scandia stabilization are all real materials, as are rare-earth zirconates and multilayer architectures that reduce yttrium content. None is a drop-in. Each changes phase behaviour, thermal expansion, sintering response, and failure mode.
Which brings the actual constraint into focus. Aviation qualification cycles are measured in years rather than quarters, which is why credible assessments place meaningful aerospace demand reduction at 2027 or later. Semiconductor chamber-part requalification at advanced nodes is a multi-quarter exercise, and available alternatives typically erode faster or generate more particles. For SOFC, an electrolyte change propagates into sintering schedules, thermal expansion matching against electrodes and interconnects, and seal design — swapping the electrolyte is a stack redesign, not a substitution, which is the context for how we scope SOFC materials work.
So the honest summary of sixteen months is this: for any qualified application, the question was never whether an alternative exists. It was whether the buyer could afford the requalification window. Almost nobody could. So record prices changed purchasing behaviour a great deal — rationing, idled lines, grade downshifting where the property bought was mechanical — and changed almost none of the qualified material specifications.
What This Predicts for the Next Round
The October 2025 tranche — MOFCOM Notices 55 through 62, including extraterritorial controls reaching foreign-made products containing 0.1% or more Chinese-origin rare earth content by value, assessed at component rather than finished-assembly level — was suspended on November 7, 2025 and that suspension runs to November 10, 2026. It re-imposes then unless extended. As of early August 2026 that is roughly three months away, and it is the single date a YSZ buyer should have in a calendar. The full sequence of Chinese materials controls is tracked in our export-controls timeline.
If it re-imposes, the change for YSZ is qualitative rather than incremental. Today the sintered-product gray zone is an origin question: was this article exported from China? A content threshold assessed at component level converts it into a lineage question: does this coated part or this electrolyte layer contain Chinese-origin yttrium above the threshold, wherever it was made? That is a much harder question, because answering it requires knowing your supplier's feedstock provenance rather than your own shipping documents.
Three inferences carry forward from this round:
- Assume inclusion, then argue out of it. Controls are written at the element and oxide level and land unevenly by tier. The absence of a concentration floor in the yttrium listing showed that low content is a weak defence.
- The gray zone will not be clarified for you. Sixteen months produced no published guidance on sintered ceramics. Design contracts and inventory around persistent ambiguity rather than waiting for a ruling.
- Feedstock lineage becomes a contractual item. If the next threshold is content-based, "where did your yttria come from" is a question you need answerable on demand — which means asking it now, in writing, while it is still merely prudent. The broader licensing-era playbook is in procurement strategy under export controls.
Frequently Asked Questions
Are the yttrium export controls still in force in August 2026?
Yes. Announcement No. 18 of April 4, 2025 remains in force and was never suspended. The measures suspended in November 2025 were the October 2025 additions, and that suspension expires November 10, 2026. Yttrium exports still require a MOFCOM licence per shipment unless the exporter holds a general licence.
Is yttrium the only one of the seven controlled items that matters for YSZ?
For YSZ specifically, yes — yttria is the stabilizing additive in every YSZ grade. Scandium matters for the related ScSZ electrolyte used in some SOFC designs, which we cover separately. The other five controlled items sit in magnet, imaging, and scintillator supply chains rather than in stabilized zirconia.
Why did record yttrium prices not force substitution?
Because the constraint is qualification time, not price. Alternative stabilizers and non-Chinese yttria both exist, but aviation qualification runs in years and semiconductor chamber requalification in multiple quarters, and reclaimed or new-source powder faces the same gate. Where a buyer cannot absorb that window, paying the higher price is the rational choice.
What actually changes on November 10, 2026?
If the suspension is not extended, the October 2025 extraterritorial provisions return, including a 0.1%-by-value Chinese-origin rare earth threshold assessed at the component level. For YSZ that would shift the question from where an article was exported to whose yttrium is inside it, which requires visibility into a supplier's feedstock provenance.
Has anyone published how many yttrium export licences are approved?
Not that we can verify. Review times have been described informally as 45 to 60 days with observed cases running longer, and MOFCOM confirmed in December 2025 that some general licences had been approved without publishing counts or terms. No yttrium-specific approval or refusal rate has been released, so any percentage figure should be treated as an estimate.
References (Public Sources)
- China MOFCOM and General Administration of Customs, Announcement No. 18 (April 4, 2025)
- Holland & Knight; CIRS Group — analysis of China's medium and heavy rare earth controls
- Center for Security and Emerging Technology (CSET) — translation of MOFCOM Notice 2025 No. 61
- White & Case — extraterritorial jurisdiction and content thresholds in the October 2025 measures
- Pillsbury; Clark Hill — analysis of the November 2025 suspension of Notices 55–62
- CGTN; South China Morning Post — MOFCOM confirmation of first general licences (December 2025)
- CMS; Certivo — licence application requirements and observed review periods
- Silverado Policy Accelerator — China rare earth export flows, June 2026 update
- The Oregon Group — yttrium oxide price movement and export volume decline
- Shanghai Metals Market, via rare-earth-mining.com — China domestic yttrium metal benchmark, August 2026
- Materials Dispatch — yttrium downstream impact, US import volumes, and qualification timelines
- USGS Mineral Commodity Summaries — yttrium and rare earth production estimates
- Journal of Material Cycles and Waste Management; Journal of Environmental Chemical Engineering — YSZ and Y/Zr recovery research
- Lynas Rare Earths; MP Materials; ReElement Technologies — public capacity statements
Sixteen months of licensing produced no guidance on sintered ceramics, so feedstock lineage is something a buyer has to be able to answer in writing rather than something the rules will settle. Nami Tech Solutions (NTS) puts the yttria sourcing path and the licensing record to mills directly, in their own language, and settles origin documentation, classification and the licence contingency in the contract as ordinary trade practice.
