25 September 2026 · 07:00 ET
Issue 01
Four thousand tonnes
- The global market for dysprosium and terbium oxide is on the order of 4,000 tonnes a year KR, derived from a developer's disclosure that its 600 tonne per year plant will represent approximately 15% of world output P and cross-checked against a second developer's phase-one tonnages. The IEA puts the downstream production those elements gate at US$6.5 trillion a year P. That ratio is the whole argument of this note.
- The bottleneck is not the mine. On IEA's stage-by-stage data China holds 60% of mining, 91% of refining and separation, and 94% of sintered magnet production, and ex-China capacity counting every announced project meets roughly half of 2035 mining demand but well under a fifth of magnet demand P. The market prices this backwards.
- Two allied governments have independently converged on the same administered price. The US Department of Defense floor for NdPr is $110/kg, and the Japanese state vehicle's upgraded Lynas agreement of March 2026 carries the same $110/kg floor S. Any project economics still marked against a $60 to $75 long-run deck is marked against the wrong curve.
- The floor is already non-binding where it matters. North American assessments launched at $125/kg for NdPr, above the floor, and at $1,200/kg for dysprosium and $4,900/kg for terbium against Chinese-basis 2025 averages of $239 and $1,010 PS. The security-of-supply premium is roughly 1.7 times for light rare earths and about five times for the heavies.
- We expect three separate Chinese control instruments to reach expiry between 10 November and December 2026, including the 0.1% by-value extraterritorial rule, which on a $40,000 vehicle is triggered by $40 of controlled content P. Buyers should be treating November as a compliance event, not a trade-policy headline.
Size the market before you size the opportunity
Almost every projection in this sector starts from demand growth and works down. Start instead from the size of the thing being fought over, and the strategic picture changes shape.
Carester's Lacq project states that it will produce 600 tonnes a year of dysprosium and terbium oxides, approximately 15% of global production P. That implies world output of roughly 4,000 tonnes a year KR. Energy Fuels' phase one at White Mesa targets approximately 20 tonnes of terbium and 120 tonnes of dysprosium, or 140 tonnes, which is 3.5% of that total P. Two independent developers, two disclosures, one consistent scale.
At North American assessment levels that entire market is worth somewhere between three and five billion dollars a year. It is smaller than the revenue of a single mid-cap miner, and it is a precondition for the traction motors, wind generators, actuators and defense platforms that the IEA associates with US$6.5 trillion of annual production value P.
A floor at $110/kg NdPr is now non-binding for light rare earths outside China, because the Western clearing price is above it, and irrelevant for the heavies, which trade at roughly five times the Chinese basis with essentially no qualified non-Chinese supply. The marginal project question in 2026 and 2027 is not whether a developer can clear $110 on NdPr. It is whether anyone can separate dysprosium and terbium at all. Nearly all announced Western capital is pointed at light rare earth mining.
Two governments, one number, two different deals
In July 2025 the US Department of Defense agreed a contract for difference with MP Materials at $110/kg NdPr. In March 2026 the Japanese state vehicle JARE upgraded its Lynas agreement to 2038, with a floor volume of at least 5,000 tonnes a year of PrNd at a floor price of $110/kg, a profit share triggered only above $150/kg, a 30% share of the excess above that trigger, and an annual repayment cap of $10 million S.
The identical headline number is not the interesting part. The structures differ in a way that is directly usable as a negotiating benchmark. The DoD clawback bites immediately at the floor. The Japanese deal has a $40 dead band before any clawback starts and caps the clawback in absolute terms. Japan's structure is materially more generous to the producer on the upside, and any producer signing government offtake in the next twelve months should be asking for the Japanese shape rather than the American one.
The cost of the American floor is disclosed quarterly and is falling fast: $51.0 million, then $42.3 million, then $17.6 million across the first three quarters of the agreement P. That is a 65% decline in two quarters and it is the cleanest public read anyone has on the direction of the benchmark price MP actually realises, because it is a filed number rather than an assessment. On our reconstruction the benchmark has closed roughly three-quarters of the gap to the floor since the agreement began; on that trajectory the floor stops paying out in late 2026 or the first half of 2027, at which point the clawback begins running in the government's favour KR.
What the customs data actually says, and what it does not
Chinese rare earth exports on the aggregate customs category ran 30,482.8 tonnes in the first half of 2026, down 6.4%, on a value of US$308.3 million, up 61.1% S. January to July came to 34,706.3 tonnes, down 10.0% S.
Do not read the value increase as a price signal.
The implied average unit value moves from US$5.87/kg to US$10.12/kg KR, which is irreconcilable with any magnet or NdPr price. This aggregate category is dominated by cheap lanthanum and cerium compounds, so a 72% rise in unit value on a 6.4% fall in tonnage is a composition change, not a price move. Anyone reading "rare earth export value up 61%" as evidence of demand strength is misreading a mix shift. This is precisely the error a customs cross-check exists to catch.
The finished magnet series tells a different and more useful story. July 2026 magnet exports were 5,375.1 tonnes, down 3.6% year on year, with the EU at a four-month high, the United States at a record July, and Japan at 111.4 tonnes, down 52.1%, the lowest since May 2025 S. Every other major destination grew. August rose 10.2% month on month S.
Two readings are available for the Japanese collapse: deliberate Chinese allocation away from Japan, or Japanese substitution finally landing through the Lynas channel. Given that the JARE agreement was upgraded five weeks earlier and now directs 50% to 75% of Lynas heavy rare earth output to Japan, we think substitution is the better hypothesis, and it is testable against Japanese import statistics. It is the most under-reported number in the 2026 trade data.
The wider pattern in the same data is the shape of the policy itself. Finished magnet exports have largely returned to pre-control levels while exports of rare earth compounds and metals remain well below historical levels S. China is willing to export magnets and not the inputs to make them. That is an industrial policy operating as designed, not a sanction.
| Destination | Tonnes | Y/y |
|---|---|---|
| European Union | 2,141.9 | +0.1% |
| United States | 647.3 | +4.5% |
| South Korea | 606.5 | +19.3% |
| Japan | 111.4 | −52.1% |
| Russia | 84.3 | +1,026% |
| Total | 5,375.1 | −3.6% |
Metallization is the stage nobody funded
Between separated oxide and a finished magnet sits oxide-to-metal conversion, a fluoride-based molten salt electrolysis step run at around 1,050 degrees Celsius requiring specialised refractories and scarce metallurgical labour. China controls approximately 90% of it S.
The scale comparison is the one to put in front of a sourcing committee. China Northern Rare Earth alone budgeted approximately 45,000 tonnes of NdPr and Nd metal production for 2025. Total announced ex-China metals, alloys and finished magnet capacity is approximately 18,000 tonnes S. One Chinese company's single-year metal budget is two and a half times the entire announced Western pipeline.
Announced nameplate is not qualified output, and the gap is measurable.
MP began magnet production in December 2025 and was still describing deliveries for customer qualification and regulatory testing in its August 2026 release, eight months later P. USA Rare Earth commissioned phase 1a in March 2026 and guides its 600 tonne run rate only for end-Q4 2026, nine months from commissioning to nameplate, on its own numbers P. Australian Strategic Materials has 1,300 tonnes per annum installed in Korea and dispatched approximately 45 tonnes in the second quarter of 2026, a run rate near 14% of installed capacity SKR. And USGS, on heavy rare earth processing specifically, states that at least five companies were developing commercial-scale capability in 2025 and that none produced sustained commercial-scale quantities P.
Our working rule for clients, and we would rather state it than imply it: treat announced nameplate as reaching qualified, automotive- or defense-accepted output roughly 24 to 36 months after first production, and discount 2028 nameplate by 60% to 80% for planning purposes KR. Commissioning a press line is not the milestone that matters. Qualifying a magnet into a traction motor or a defense platform is a multi-year process and it is the one that gates revenue.
Three numbers this industry repeats that we will not print
Rare earth content per F-35. The universally quoted figure of approximately 417 kilograms traces to a single 2012 Department of Defense internal study on recycling feasibility that was submitted to Congress and never publicly released, cited once in a 2013 Congressional Research Service report U. There is no published methodology and no bill of materials. It has been repeated for thirteen years without verification, and the companion figure for a Virginia-class submarine has the same provenance. Any defense exposure model built on it should be rebuilt.
Humanoid robot magnet intensity. The demand equation is robot volumes multiplied by actuators per robot multiplied by magnet loading. We searched for a sourced value for each of the three variables and found none U. Charts projecting robot-driven rare earth demand are built on invented inputs, and we would rather say so.
Chinese production quotas. The 2024 quotas were published at 270,000 tonnes for mining and 254,000 for smelting and separation. The 2025 quotas were issued without public announcement and recipients were instructed not to disclose them, and no 2026 figures are public S. A series analysts have relied on since 2006 has gone dark. Any circulating 2025 or 2026 quota number is unsourced. The practical consequence is that customs data has become the only remaining independent read on Chinese supply behaviour, which is why we build on it.
Position summary
Size first. A roughly 4,000 tonne per year heavy oxide market gates trillions of dollars of downstream output.
The bottleneck ranks magnets, metal, separation, mining. Capital is flowing in the reverse order.
$110/kg is now an allied reference price, not a US subsidy, and it is already below the Western clearing level.
November is a compliance event. Three instruments expire in seven weeks and the 0.1% rule is the one that matters.
Discount nameplate by 60% to 80% for 2028 planning.
| Oxide | China basis 2025 avg | CIF N. America 31 Mar 2026 |
Ratio |
|---|---|---|---|
| Neodymium–praseodymium | 73–74 | 125 | 1.7× |
| Dysprosium | 239 | 1,200 | 5.0× |
| Terbium | 1,010 | 4,900 | 4.9× |
| Date | Instrument |
|---|---|
| 10 Nov 2026 | Suspensions of Announcements 55, 56, 57, 58, 61 and 62 expire, including the 0.1% extraterritorial de minimis rule |
| 27 Nov 2026 | Separate suspension of Announcement 46 Article 2 expires, covering gallium, germanium and antimony |
| ~Dec 2026 | One-year general licences issued to three magnet makers reach expiry |
| In force | Announcement 18 licensing on seven medium and heavy rare earths has never been suspended |
How to read our figures
Every material number in this note carries a provenance tag. We do not present sourced data and our own estimates as the same thing.
| P | Confirmed at a primary source we name |
| S | Secondary reporting of primary data; reporter named |
| KR | Kastra Research calculation; method stated |
| U | Circulating but unverified; do not rely on it |
What we could not verify
Any 2026 Myanmar export volume or value. No source we found carries 2026 data.
Chinese export licence approval rates. No official statistic exists; the 45-day statutory window is documented, actual performance is anecdotal.
The absolute August 2026 magnet export tonnage. We have the percentage change only.
Recycling cost per kilogram against primary separated oxide. We are not publishing a recycling cost curve this edition.
Status of one US-funded heavy separation facility, unaddressed in its parent's full-year results despite roughly $258m of allocated support.
Politics and policy
This is a market where the policy instrument is the product. Roughly nothing in the pricing of heavy rare earths is explicable without reference to four governments acting deliberately, and two of those four have published the terms on which they will act. Our conviction on political outcomes is graded lower than on the market analysis above.
- China
Export control as industrial policy, not as sanction
The pattern in the 2026 trade data is consistent and it is the whole story: finished magnet exports have largely returned to pre-control levels while exports of rare earth compounds and metals remain well below historical levels S. China is willing to sell the product and not the capability to make it.
That distinction matters for how the restriction is read. A sanction is leverage applied to change a counterparty's behaviour, and it lifts when the behaviour changes. An industrial policy protects a domestic value-added position and it does not lift, because there is nothing to concede.
Read. Do not model the controls as a negotiable item. Analysts waiting for a trade deal to restore compound and metal flows are waiting for a concession that serves no Chinese objective. The controls are doing exactly what they were designed to do, which is why magnet exports were allowed to recover.
- United States
The price floor is self-liquidating, and that is a political fact
The Department of Defense contract for difference with MP Materials sets a floor at $110/kg NdPr, with a clawback that bites immediately above it. The disclosed quarterly cost has run $51.0 million, then $42.3 million, then $17.6 million P, a 65% decline in two quarters.
On our reconstruction the benchmark has closed roughly three-quarters of the gap to the floor, and on that trajectory the floor stops paying out in late 2026 or the first half of 2027, at which point the clawback runs in the government's favour KR.
Read. A subsidy that turns into a revenue line before the next appropriations cycle is a subsidy that survives a change of administration. We think the political durability of the US floor is underestimated, and that the risk to producers is not withdrawal of support but the clawback terms on which support was accepted.
- Japan
The more producer-friendly template, and the one to negotiate against
The JARE upgrade to the Lynas agreement runs to 2038 with a floor volume of at least 5,000 tonnes a year of PrNd at a floor price of $110/kg, a profit share triggered only above $150/kg, a 30% share of the excess above the trigger, and an annual repayment cap of $10 million S. The headline number is identical to the American one; the $40 dead band and the capped clawback are not.
The agreement also directs 50% to 75% of Lynas heavy rare earth output to Japan, which is the likeliest explanation for Japanese magnet imports from China falling to 111.4 tonnes in July 2026, down 52.1%, while every other major destination grew S.
Read. Any producer signing government offtake in the next twelve months should be asking for the Japanese shape rather than the American one, and should say so using these published terms. The most consequential political event in this market in 2026 was a financing structure, not a tariff.
- Defence procurement
A policy edifice resting on an unpublished 2012 study
The rare earth content of an F-35, universally quoted at approximately 417 kilograms, traces to a single 2012 Department of Defense internal study on recycling feasibility that was submitted to Congress and never publicly released, cited once in a 2013 Congressional Research Service report U. There is no published methodology and no bill of materials, and the companion figure for a Virginia-class submarine has the same provenance.
That number is nonetheless load-bearing in appropriations argument, in stockpile sizing, and in a good deal of sell-side defence exposure work.
Read. Political support built on an unverifiable figure is politically fragile, because a single credible audit can reverse it. We would not underwrite a defence demand case, or a policy-durability case, on the 417 kilogram figure, and we flag that any client model containing it should be rebuilt.
Key debates
-
Has the export-control regime effectively ended?
Six announcements were suspended in November 2025, magnet exports have returned to roughly pre-control levels, and general licences were issued to major producers.
Our answer. No, and the misreading is specific. Announcement 18, the licensing regime covering samarium, gadolinium, terbium, dysprosium, lutetium, scandium and yttrium, was never suspended P. What was suspended was the equipment, technology and extraterritorial layer, and those suspensions expire on 10 November 2026. Magnets flow; compounds and metals do not. That is the regime working, not lapsing.
-
Is the $110/kg floor good policy?
It has cost the US government $110.9 million across three quarters and the market price has risen to meet it, which looks like success.
Our answer. It solved the wrong problem cheaply. The floor de-risks light rare earth production at a moment when the Western clearing price for NdPr already sits above it. It does nothing for dysprosium and terbium, where the ex-China premium is five times and where USGS says no one has yet produced at sustained commercial scale. The Japanese structure is instructive here too: it secured element-level allocation rights, a percentage of heavy output rather than a tonnage. That is the variable that has predicted success, not the dollar amount.
-
Will announced Western capacity close the gap by 2030?
Roughly 30,000 tonnes a year of ex-China magnet nameplate has been announced for 2028 and 2029, alongside separation projects in Australia, France, Utah and Estonia.
Our answer. Not on the announced schedule, and the evidence is in the ramps rather than the announcements. Eight months from first production to still-in-qualification at one producer; nine months from commissioning to nameplate on another's own guidance; 14% utilisation at a third. IEA's own arithmetic has ex-China capacity meeting well below 20% of 2035 magnet demand counting everything announced. We would plan on that, and treat any individual project's 2028 date as a 2030 date.
-
Is recycling a real supply source or a story?
Under 5% of rare earth magnets are recovered globally, and operating magnet recycling capacity worldwide is on the order of 300 to 500 tonnes a year of magnet feed.
Our answer. Small, but structurally distinctive, and the reason is specific: recycled feed is the only Western source that is naturally dysprosium and terbium bearing without building a heavy separation circuit. Announced 2027 and 2028 capacity of roughly 5,000 to 6,000 tonnes of feed converts to perhaps 1,500 to 1,800 tonnes of mixed oxide, comparable to one primary project, but with the heavies already in the mix. We are not publishing a cost curve because no defensible cost comparison exists.
What would change our mind
- MP Materials reporting price protection income at or near zero for Q3 2026. That confirms the benchmark has crossed $110/kg and moves the floor from a subsidy to a clawback, changing the fiscal read on the whole programme.
- Renewal of the Chinese suspensions before 10 November 2026, particularly the 0.1% extraterritorial rule. Non-renewal is our base case and a renewal would defer the compliance event by a year.
- Chinese magnet exports to Japan recovering above 200 tonnes a month while Lynas heavy output to Japan holds. That would break our substitution reading of the July collapse and point to allocation politics instead.
- Any producer reporting sustained commercial-scale separated dysprosium or terbium output outside China. USGS's finding that nobody has yet done this is the single load-bearing fact in our heavy rare earth thesis.
- A qualified automotive or defense magnet award to a Western producer with disclosed tonnage. That would compress our 24 to 36 month qualification rule and make announced 2028 nameplate more credible.
- Publication of Chinese 2026 mining and separation quotas. The return of that series would reduce the analytical value of the customs cross-check we build on.
Watch list
| Date | Event | Why it matters |
|---|---|---|
| Monthly | China customs preliminary and detailed releases, around the 7th and 20th | Our primary cross-check, and the only remaining independent read on Chinese supply behaviour since the quota series went dark |
| Early Nov 2026 | MP Materials third quarter results | Next price protection print; the cleanest read on whether the benchmark has crossed the floor |
| 10 Nov 2026 | Chinese suspensions of Announcements 55, 56, 57, 58, 61 and 62 expire | The 0.1% by-value extraterritorial rule and the technology and personnel controls snap back absent renewal |
| 27 Nov 2026 | Separate suspension of Announcement 46 Article 2 expires | Different date, covering gallium, germanium and antimony, and almost universally missed |
| ~Dec 2026 | One-year general licences held by three Chinese magnet makers reach expiry | Renewal or non-renewal is the operational signal, more informative than the announcement expiries |
| Late 2026 | Caremag Lacq start-up, 600 t/yr of dysprosium and terbium oxide | The single largest addition to non-Chinese heavy supply in the window, roughly 15% of world output |
| End Q4 2026 | USA Rare Earth Stillwater reaches its 600 t/yr run rate, on company guidance | First hard test of Western magnet ramp guidance against our discount rule |
| 1 Jan 2027 | US defense sourcing restrictions extend to mined materials | Brings NdFeB and SmCo magnets and NdPr oxides and metals into scope for covered nations |
| Q4 2027 | Energy Fuels terbium and dysprosium circuits commissioned at White Mesa | First US commercial heavy separation |
| 2028 | MP Materials 10X Northlake commissioning commences | Ten thousand tonne platform under a ten-year government offtake |
Method and data sourcing
Production, reserve, import-reliance and annual average price data are from the USGS Mineral Commodity Summaries 2026 rare earths and rare earths (heavy) chapters. Stage-by-stage capacity shares and 2035 adequacy figures are from the International Energy Agency's Rare Earth Elements report, 2026.
Trade volumes are Chinese customs data as reported by named news and trade outlets. We have not queried the customs database directly this cycle and all such figures are tagged as secondary accordingly. Implied unit values are Blue Kastra calculations dividing reported value by reported tonnage and are shown only to demonstrate that the aggregate category is compound-dominated.
Company operating and financial data are from securities filings and company releases as dated in the text. Price protection agreement income is a disclosed line item; the implied benchmark price is a Kastra Research reconstruction, because the volume base to which the agreement applies is not disclosed.
Price levels are drawn from two different bases: USGS annual averages on a Chinese domestic basis, and North American assessment launch levels. The ratio between them is presented as an order of magnitude for the security-of-supply premium and not as a tradeable spread.
We have deliberately not published rare earth content per defense platform, humanoid robot magnet intensity, Chinese 2025 or 2026 production quotas, export licence approval rates, or a recycling cost curve, because no primary source for any of them exists.
Exhibits numbered 1, 2, 3 run in the body of the note; exhibits lettered A, B, C sit in the margin column. Each series is in the order it is meant to be read.
Notes
- 1USGS Mineral Commodity Summaries 2026 reports US net import reliance for rare earth compounds and metals as 67% for 2025, and elsewhere references 53% for 2024. We have not reconciled the two and have used neither in a calculation.
- 2China's mining share is variously given as 60% on a magnet rare earth basis, 61% on all mined supply, and 69% on the arithmetic of the USGS production table. These are different denominators rather than contradictions. We use the IEA magnet rare earth basis throughout.
- 3The terms of the March 2026 Japanese supply agreement are drawn from trade reporting rather than from the underlying exchange announcement and are tagged as secondary throughout. This is the most consequential finding in the note and carries the weakest sourcing.
- 4The August 2026 magnet export month-on-month change is reported; the absolute tonnage and destination detail were not available at our data cutoff.
Principal sources
USGS, Mineral Commodity Summaries 2026, rare earths and rare earths (heavy) chapters. International Energy Agency, Rare Earth Elements, 2026. MP Materials quarterly earnings releases via SEC EDGAR, 26 February, 7 May and 6 August 2026, and corporate releases of 10 July 2025, 15 July 2025, 19 November 2025 and 26 February 2026. Lynas Rare Earths FY26 results, 26 August 2026. JOGMEC release, 7 March 2023. METI release, 17 March 2025. Carester, 17 March 2025. Neo Performance Materials, 22 September 2025. USA Rare Earth, 26 March 2026. Energy Fuels, 29 July 2026. Vulcan Elements, 3 November 2025. Noveon Magnetics, 19 January 2026. European Commission Critical Raw Materials Act strategic project lists, March and June 2025. China General Administration of Customs data as reported by Global Times, ANI, Reuters and specialist trade services, and Ministry of Commerce announcement numbers as reported in independent legal summaries. S&P Global Platts rare earth assessment launch, 31 March 2026.
Abbreviations
- REO rare earth oxide
- NdPr neodymium–praseodymium
- Dy dysprosium
- Tb terbium
- Sm samarium
- HREE heavy rare earth elements
- NdFeB neodymium iron boron magnet
- SmCo samarium cobalt magnet
- HS 8505.11 customs code, permanent magnets of metal
- CIF cost, insurance and freight
- CRMA EU Critical Raw Materials Act
- MOFCOM Ministry of Commerce, China
- IEA International Energy Agency
- t/yr tonnes per year
Important disclosures
About this publication. Kastra Research is an independent research publisher. We are not a broker-dealer, investment adviser, or registered investment research firm. We do not underwrite, trade, or make markets in any security or commodity, we publish no ratings or price targets on securities, and we make no recommendation to buy, hold, or sell any security.
Analyst certification. The author certifies that the views expressed in this report accurately reflect their personal views about the subject commodities, markets, and entities, and that no part of their compensation was, is, or will be directly or indirectly related to the specific views or conclusions expressed.
Intended recipients. This report is prepared for institutional and professional recipients with the resources and expertise to make their own procurement and investment decisions. It is not directed at retail investors, is not a personal recommendation, and takes no account of the objectives, financial situation, or particular needs of any recipient.
No reliance. Information in this report is drawn from sources we believe to be reliable, but we make no representation or warranty, express or implied, as to its accuracy, completeness, or timeliness, and none should be inferred. The provenance tags exist precisely because reliability is not uniform across the figures we print. Nothing here is a substitute for a recipient's own verification, and no recipient should act on any figure, estimate, or conclusion in this report without independently confirming it against the underlying source, which we name so that they can.
Forward-looking statements. This report contains estimates and scenario analysis. Forecasts depend on the assumptions stated alongside them and actual outcomes will differ, potentially materially. Historical relationships are not a reliable indicator of future results.
Political and policy commentary. This report discusses legislation, regulation, litigation, sanctions and trade measures because they are material to the markets covered. That discussion is analysis of market consequence, not political advocacy, and it is neither an endorsement of nor opposition to any government, party, candidate, official, or policy position. Where we describe what a political actor is likely to do, we are forecasting behaviour in order to price risk. The forward-looking caveat above applies to those passages with additional force: political outcomes are less tractable than industrial ones, and our conviction is graded lower in that section for that reason.
Use of artificial intelligence. This report was produced with the assistance of artificial intelligence. AI tools were used in gathering and summarising source material, in drafting and editing the text, and in preparing the exhibits and the calculations behind them. AI systems misattribute sources, transpose figures, and state false things fluently, so the provenance discipline in this note is the control on that failure mode rather than a presentational device: every figure carrying a P or S tag was checked by the author against the named source, and figures the author could not confirm are tagged U or listed under "What we could not verify" rather than presented as established. The judgements, the conviction levels, and any errors are the author's. The use of AI does not qualify the analyst certification above and does not reduce the author's responsibility for anything in this document.
Conflicts. Kastra Research receives no compensation from any entity discussed in this report. No subject entity reviewed this report prior to publication, and we do not accept payment for coverage.
Jurisdiction and distribution. This report is published from the United States and is not directed at any person in any jurisdiction where its publication, distribution, or use would be contrary to law or regulation, or would subject Kastra Research to any registration or licensing requirement. It is not an offer to sell, or a solicitation of an offer to buy, any security, commodity, or instrument in any jurisdiction. Recipients are responsible for compliance with the laws applicable to them, and for observing the redistribution terms below.
Limitation of liability. To the fullest extent permitted by law, Kastra Research, its owners, and its contributors accept no liability for any loss or damage, whether direct, indirect, incidental, consequential, or otherwise, arising from any use of or reliance on this report or any part of it. Recipients use this report at their own risk.
© 2026 Kastra Research. Licensed to the named subscriber. Redistribution in whole or in part requires prior written consent. Quotation is permitted with attribution.