Rare earths after Greenland: what can $CRML, $MP, $USAR and $UUUU actually sell?

01The Greenland headline now has a text—and a separate effectiveness condition

The useful question about rare earths is becoming more precise: what can a company deliver, in what chemical form, from which facility, and under which commercial terms? Four tickers can respond to the same geopolitical headline while representing very different answers. Critical Metals brings a Greenland development project and a proposed processing route. MP Materials combines established mining and separation operations with a growing downstream business. USA Rare Earth has just changed its perimeter through Serra Verde. Energy Fuels combines uranium, rare-earth separation and further downstream expansion. Their shared sector label does not make their present products interchangeable.

On September 22, 2026, the United States, Denmark and Greenland signed an agreement amending and supplementing the 1951 defense framework. The full text matters more than a social-media description. It explicitly includes resource extraction among sensitive activities and sets investment-screening provisions affecting control, significant influence and access to sensitive information by investors from countries outside NATO, NATO partnerships or the European Union, subject to the agreement’s qualifications. Implementation sits with Greenland in cooperation with Danish authorities. The document also reaffirms sovereignty and Greenlandic self-determination. Read the signed agreement.

There is a second date to watch. Article XII makes entry into force conditional on a diplomatic note confirming completion of the necessary parliamentary procedures. The September 22 signature therefore should not be casually relabeled as evidence that every provision is already effective. This research confirms the signed text and its stated condition; it does not establish that the subsequent diplomatic notification has occurred. The Danish prime minister’s office independently documents the agreement and the participation of the Greenlandic government at the September 22 press conference. Danish government account.

For shareholders, the immediate analytical task is to connect that framework to an actual corporate right. The agreement does not itself award CRML a new mine, transfer ownership of an asset to one of these four companies, establish a named customer order for their products or fund a specific corporate construction budget. Those would require separate documents. Investment screening can affect which counterparties are acceptable; an acceptable counterparty still needs a contract, technical capability and money. This distinction is our reading of the agreement’s scope, rather than a prediction of the outcome of any particular application.

The gap is visible in the discussion itself. A sampled Reddit thread asks whether the Greenland development is an agreement, a completed signing or a revival of an older framework. It is useful as evidence of reader uncertainty, not as a representative measure of investor sentiment or as confirmation of the share-price figures quoted by individual posters. The official text answers the documentary question. Company filings must answer the operating one. Discussion that illustrates the confusion.

02A tonne of rare earths is not a tonne of finished magnets

The chain runs through several separate products: mined material, concentrate or an intermediate chemical product, separated oxides, metals, alloys and finished magnets. The IEA’s 2026 analysis explicitly examines these stages rather than treating mine capacity as a substitute for the whole chain. Its magnet-focused rare-earth grouping consists of neodymium, praseodymium, dysprosium and terbium. A total rare-earth figure can contain a much broader basket. IEA value-chain illustration, IEA executive summary.

For this comparison, every capacity figure needs a product label. A tonne of mineral concentrate contains gangue and multiple constituents; it is not a tonne of separated neodymium-praseodymium oxide. An oxide includes oxygen, so even a pure oxide quantity cannot be carried unchanged into a metal calculation. A finished magnet contains an engineered composition and has its own processing losses. Reporting all three as simply “tonnes of rare earths” can make businesses look far closer in scale than the source documents justify.

The location of the missing step also matters. A company with an established mine may need a customer to perform separation. A separator may depend on purchased material whose composition varies. A magnet facility may need a particular alloy even when the corporate group owns an upstream deposit. Vertical integration is valuable when the links work together; common ownership alone does not remove transport, qualification, scheduling or working-capital requirements. An acquisition can move a box on an organizational chart immediately while the material flow takes longer to integrate.

Purity and recovery answer different questions. Recovery describes how much of a starting component is recovered through a defined process; purity describes the composition of an output stream. A very pure sample can be produced at a low yield. A strong dissolution result can leave separation and finishing unproven. An industrial assessment therefore needs the input, the output, the denominator and the process boundary. It also needs to establish whether the result came from a selected sample, a pilot campaign or sustained commercial operations. The company sections below keep those boundaries attached to the figures.

Consider a simple, entirely hypothetical mass balance. A plant receives 100,000 tonnes of feed containing 2% total rare-earth oxides on the stated assay basis. The starting inventory is 2,000 tonnes of contained oxide equivalent. At an assumed 80% recovery through the first stage, 1,600 tonnes remain on that same equivalent basis. If a target component represents 20% of that recovered basket, its inventory is 320 tonnes; a further 90% recovery yields 288 tonnes of target oxide equivalent. None of these quantities is a finished-magnet tonnage, and none is an estimate for the four companies.

This example is useful because the denominators remain visible. If first-stage recovery falls to 70% with all other assumptions unchanged, the result becomes 252 tonnes, a reduction of 36 tonnes. A 99% figure quoted for one intermediate step would not erase that loss elsewhere. Nor would an attractive selling price rescue every composition: the quantity of the desired component, the cost of recovering it and the amount a customer actually pays for the delivered intermediate all matter. A payable percentage belongs in the commercial calculation after the physical balance; it should not be silently treated as an additional extraction recovery.

03Geological scale, industrial proof and the budget to connect them

A large deposit establishes a potentially valuable starting point, but the classification used in a technical report is part of the fact. The SEC’s mining definitions distinguish resources from reserves, and distinguish geological confidence from the technical and economic work required to establish an economically mineable quantity. Inferred resources carry a lower confidence level and cannot simply be relabeled as reserves. A feasibility study also covers much more than a laboratory result, including design, operating assumptions and the relevant modifying factors. SEC mining definitions, Item 1300.

An investor can apply that distinction without becoming a geologist. Start with the document’s actual category and date. Then identify the processing route assumed, the annual feed rate, the product sold and the capital estimate. Finally ask which of those assumptions has since been validated or changed. A later press release may improve one part of the project without replacing the entire earlier technical assessment. Adding new results to an old production schedule without a formal reconciliation can create an investment case that no source actually presents.

There are at least three different commercial milestones in a processing project. Producing an acceptable sample demonstrates a result for that sample. Customer qualification demonstrates acceptance under a defined specification and procedure. Recurring invoiced shipments show that the supplier can deliver under commercial conditions. Each achievement can reduce uncertainty, but the documents should state which one occurred. A shipment for testing or regulatory approval should not be counted automatically as evidence of a mature revenue stream.

The financing bridge needs a similarly explicit boundary. A capital estimate measures expected expenditure for a defined project scope; a commitment identifies a potential funding source; cash on the balance sheet measures funds already present at a date. They are related but cannot be substituted for one another. Committed financing can have conditions, expiry dates, collateral, milestones and future fundraising requirements. Acquiring an operating asset can add revenue while also adding debt, integration costs and share issuance. What matters is the combined business after those obligations, not the size of a transaction headline.

This is the framework behind the next four company analyses. The comparison is intentionally organized around product, evidence and remaining work. It does not rank the shares or assign a common valuation multiple to activities at different stages. Where the only direct operational evidence is an issuer statement, it remains labeled as company-reported information; a second document is used when it adds a relevant check rather than merely repeating the announcement.

04CRML: owning Tanbreez and financing its development are separate milestones

Critical Metals offers exposure to a Greenland rare-earth development project whose ownership has become clearer, while its industrial economics still require substantial demonstration. On April 30, 2026, the company completed the acquisition of an additional 50.5% of Tanbreez, bringing its interest to 92.5%. It issued 14.5 million ordinary shares in that transaction. European Lithium retained the other 7.5%. That is the completed ownership position, rather than the 100% position contemplated under a separate, unfinished corporate transaction. Critical Metals, April 30, 2026.

This distinction matters to the common shareholder. Acquiring more of a deposit increases the company’s economic participation, but payment in shares also increases the number of claims on that participation. The relevant comparison is the additional asset interest obtained against the shares issued and the future capital required to develop it. A larger ownership percentage, considered alone, cannot establish how much value has been created per existing share.

Tanbreez’s principal rare-earth carrier is eudialyte, within the layered kakortokite rock of southern Greenland. Its chemistry also brings zirconium, niobium and tantalum into the development case. The amended technical report, effective April 13, 2026, describes the project as being at the advanced exploration and resource delineation stage. It explicitly contains no Mineral Reserve Estimate and does not constitute an Initial Assessment under S-K 1300. The exploitation licence described in that report initially permits 500,000 tonnes a year of run-of-mine material. That is ore before concentration, not separated rare-earth production. Amended Tanbreez technical report, April 13, 2026.

The practical implication is that an attractive resource must still be translated into a mine plan, a reliable processing route and a financeable construction schedule. A geological estimate does not tell an investor how much material a customer will accept each year. Nor can a licence for an initial extraction rate support every larger downstream scenario without checking the proposed expansion and its approvals.

Physical development has progressed, but the evidence must retain its date. A June 2 update described pilot-plant foundations, infrastructure and preparations for bulk sampling. An August completion date appearing in that June material was a target, rather than subsequent proof of completion. Those activities can support the collection of better engineering data; they do not, by themselves, establish recurring commercial production. Critical Metals site update, June 2, 2026.

05CRML: what the September chemistry announcement does—and does not—establish

The September 16 announcement is potentially consequential because processing eudialyte is central to the investment case. Its headline cites more than 99% dissolution and 19 high-purity products. Its underlying description, however, concerns a proposed Romanian refinery and a preliminary process and economic model. The design contemplates up to 100,000 tonnes of concentrate feed annually, with preliminary capital expenditure of $1.85 billion, plus or minus 25%. Its approximately $2.2 billion annual revenue projection is a modeled outcome, not revenue generated or contracted at that level. Critical Metals refinery announcement, September 16, 2026.

Dissolution answers one question: whether material can be moved into solution under the tested conditions. The shareholder’s question is longer: how much of each valuable element survives the entire process, reaches the required chemical form and purity, and is sold at a price covering all costs? Losses or expensive treatment at later stages can coexist with a strong initial dissolution result. A concentrate that dissolves readily can still produce a solution that is difficult to purify selectively.

The proposed process addresses silica-gel formation through a multistage acid route and includes acid recycling, separation and finishing into individual products. These are important engineering objectives. Continuous operation must establish whether the circuit remains manageable as feed composition varies, whether impurities accumulate in recycled streams, and whether filtration and equipment availability remain acceptable. These are the appropriate tests of the proposed flowsheet, rather than conclusions that can be drawn from one percentage.

The model also places material economic weight on silica, assigning approximately $600 million of annual revenue to that coproduct. This creates an additional commercial question: which customers would absorb the specified grades and volumes at the assumed prices? High purity has value only within a market that needs the product in its delivered form. Freight, packaging, quality control and customer approval can influence the price actually realized. The source describes projected purity and process performance as requiring further validation. September 16 process and economic assumptions.

The 19-product description also includes critical metals beyond rare earths. It should not be read as 19 distinct rare-earth elements, or as 19 already qualified product businesses. Each saleable stream could broaden the opportunity, but it also brings its own specifications, inventory and market-development requirements. Our reading is that the announcement strengthens the case for further technical validation. It does not yet establish the economics of a continuously operating, fully financed refinery.

06CRML: the buyer’s version of the offtake reveals the missing commercial steps

Realloys provides a useful second primary source because it describes the agreement from the customer’s side. The definitive contract executed on May 20 covers 15% of monthly Phase 1 concentrate production, with a permitted delivery variance, and excludes later phases. The buyer describes initial nameplate output of up to 15,000 tonnes of concentrate annually. The initial 15-year term starts with the first commercial delivery. Product qualification is required, certain payables, recoveries and price floors still need finalization, and either party can terminate if initial commercial delivery has not occurred within five years. Realloys, May 21, 2026.

This is more informative than describing the agreement simply as “15% of Tanbreez sold.” It identifies a route to market for an initial project phase. It does not establish a purchase commitment over every future expansion, a present annual invoice, or unconditional payment before qualified production exists. Critical Metals’ own announcement confirms the long-term agreement, but the buyer’s more specific scope is essential to understanding it. Critical Metals, May 21, 2026.

The financial importance of “payable” material is easy to miss. A customer purchasing mixed concentrate is not buying a tonne of pure dysprosium or neodymium. The eventual invoice depends on the measured composition and the agreed portion of contained material for which the buyer pays, together with the commercial treatment of processing and other deductions. Multiplying concentrate tonnage by a pure-oxide quotation would therefore create a revenue estimate the contract does not support.

There is also a scale question to resolve. The buyer’s initial 15,000-tonne annual concentrate framework and the proposed refinery’s 100,000-tonne feed design refer to different industrial scales. This is not proof that the refinery cannot be supplied: subsequent mine expansion or other feed arrangements could change the picture. It means that investors need the phase-by-phase feed plan, timetable and financing before placing both numbers in a single operating forecast.

For the same reason, a refinery term sheet and a binding concentrate offtake should not be treated as interchangeable commitments. Our analysis would prioritize representative qualification batches, finalized pricing mechanics, a delivery schedule linked to funded construction and evidence that allocation among customers and downstream partners is consistent. A successful qualification batch would reduce one risk. Repeat accepted deliveries would reduce another. Neither should be substituted for the other when assessing the next company announcement.

07CRML: the funding question extends beyond a cash headline

The most recent complete interim statements verified here cover the six months ended December 31, 2025. They report $80.9 million of cash and cash equivalents, $19.2 million of operating cash outflow and a $120.4 million loss under IFRS. The loss includes substantial noncash warrant remeasurement and share-based expenses, so it should not be read as the amount of cash consumed. In a later, May 18 announcement, CRML reported approximately $124 million of standalone cash. That later figure is a dated company snapshot, not a September cash balance. Interim financial statements, May 18 transaction announcement.

The proposed acquisition of European Lithium could bring additional assets, liquidity and the remaining Tanbreez interest. It had not closed at the cutoff. The September 15 update scheduled shareholder and optionholder meetings for October 22, with implementation expected in November if conditions and approvals are satisfied. Consequently, the target’s cash cannot already be added to CRML’s available funds. Critical Metals, September 15, 2026.

The share exchange also deserves current terms. An August amendment replaced the original fixed exchange ratio with a variable ratio subject to a cap and collar. Existing cross-holdings complicate the net share-count effect. Investors should therefore examine final consideration, the treatment of those cross-holdings and the resulting securities outstanding, rather than equating gross issuance with net dilution—or assuming the transaction is dilution-free. Amended transaction terms, August 19, 2026.

Even a stronger combined balance sheet would leave a separate question about funding the proposed refinery. A multibillion-dollar engineering scenario needs a defined funding share, committed partners, construction contingencies and working capital. It would be premature to assume CRML must fund every dollar itself, but equally premature to assume partners will absorb those costs without obtaining economic rights. Project-level ownership and financing terms ultimately determine how much of any successful refinery belongs to common shareholders.

The most useful CRML milestones are therefore interlocking: stronger process evidence, a reconciled mine-to-refinery schedule, firm customer specifications, committed capital and an understandable post-transaction share count. Progress on these measures would make future production estimates more credible. A sequence of larger theoretical revenue figures without those connections would leave the central financing and execution questions unanswered.

08MP: separated product and customer revenue already exist, but the units need care

MP Materials begins from a different industrial position. Mountain Pass is an operating mining and processing complex; its bastnaesite-based feed supports an existing separation business. The company’s product documentation distinguishes concentrate from separated NdPr oxide, metal and downstream magnetic materials. That distinction is fundamental: having rare earths in a concentrate does not mean every contained element has already been separated into a marketable individual product. MP’s Mountain Pass history, MP product catalogue.

In the second quarter ended June 30, 2026, MP produced 840 metric tonnes of finished NdPr oxide and reported 1,006 tonnes of NdPr sales on an oxide-equivalent basis. It also produced 11,072 tonnes of rare-earth oxide content within concentrate. These figures describe different stages of the same material flow; they cannot be added into one output total. The larger sales figure does not, by itself, prove a production acceleration within the quarter. MP second-quarter results, August 6, 2026.

The sales metric also includes intersegment transactions. Metal is expressed as oxide equivalent using an assumed conversion ratio of 1.25 beginning in 2026, compared with 1.20 previously; older comparisons were not restated. A change in this convention is not a new physical tonne produced. Nor is an internal transfer an additional sale to an outside customer at the consolidated level. MP Form 10-Q, June 30, 2026.

For analysis, this suggests following three reconciliations: finished separated output, externally realized sales and inventory moving toward further processing. Stronger separation output is useful evidence that MP can execute a chemical manufacturing ramp. But its economic contribution depends on product mix, realized pricing, operating costs and whether further processing earns enough to justify the extra investment and working capital.

The same discipline applies to diversification. A catalogue can show technical breadth without disclosing a material revenue business in each listed product. It is more informative to ask whether a new separated element has been accepted by a customer, whether shipments recur, and whether management separately identifies its contribution. MP’s lead in actual operations makes those questions answerable through successive reporting periods; it does not eliminate the need to ask them.

09MP: Independence, Apple and 10X are three different commercial clocks

MP reported additional magnet deliveries from Independence during the second quarter for customer qualification and regulatory testing. At the same time, its $16.5 million of Magnetics segment revenue came entirely from magnetic precursor products. Those disclosures support genuine progress through metal and alloy toward finished magnets, while preserving an important distinction: qualification deliveries are not evidence of a mature finished-magnet sales business. MP second-quarter results, August 6, 2026.

A magnet customer is buying more than chemical purity. In practical terms, the part must fit an application, deliver the required magnetic performance and remain reliable in its intended operating conditions. A customer can accept material for testing while still evaluating whether a production process consistently meets the final requirements. Investors should therefore look for the transition from samples and qualification shipments to repeat commercial deliveries, with corresponding revenue disclosure.

Apple’s commitment provides unusually useful customer-side confirmation. In July 2025, Apple announced a $500 million multiyear purchase commitment, alongside work on dedicated manufacturing capacity and recycling. This is a commercial supply commitment, rather than a $500 million purchase of MP equity. MP identified 2027 as the expected start of magnet shipments under the partnership. Apple, July 15, 2025, MP’s partnership announcement.

Recycling adds a distinct source of feed and a customer-specific production route. The opportunity depends on gathering suitable material, recovering it efficiently and qualifying the resulting product. The existence of an agreement does not establish how much margin MP will earn per magnet or allow the entire headline commitment to be placed into a single year’s revenue. It does strengthen the commercial basis for dedicated investment because a named buyer has described the relationship itself.

The much larger 10X project operates on a later schedule. MP selected Northlake, Texas, for a campus requiring more than $1.25 billion of expected investment, with approximately $200 million in state and local incentive packages. Commissioning is expected in 2028. Its projected 7,000 tonnes of annual magnet capacity, combined with expanded Independence capacity of 3,000 tonnes, produces the frequently cited future 10,000-tonne figure. It is not current output. MP’s Northlake announcement, February 26, 2026, June 2026 Form 10-Q.

Keeping these clocks separate improves the investment analysis: Independence qualification can advance before Apple shipments begin, while 10X construction can consume cash before its own commissioning. Success at one stage supports confidence, but does not automatically bring forward revenue from the other two.

10MP: the government agreement changes price exposure, not every operating risk

MP’s price-protection agreement establishes $110 per eligible kilogram of NdPr equivalent, calculated through the contract’s quarterly mechanism. It can apply to specified stockpiled material, affiliate transfers and third-party sales, subject to the agreement’s definitions. It is not $110 for every kilogram of ore or concentrate, and it does not cover every rare-earth element. Following the specified production milestone, the contract provides the government with 30% of the relevant upside above the protected level. Price Protection Agreement, July 9, 2025.

This mechanism meaningfully reduces one source of uncertainty. It gives eligible production a different exposure to weak benchmark prices than an entirely unprotected producer would have. Nevertheless, a protected price does not guarantee an attractive margin if recovery, reagent use, labor, maintenance or other costs disappoint. A plant that fails to produce qualifying material also cannot be evaluated as though every planned tonne had received support.

Its treatment of stockpiles creates another analytical distinction. A contractual payment associated with eligible material does not necessarily mean an independent commercial customer has purchased that material. The support can improve cash economics during the domestic supply-chain buildout while leaving ordinary customer revenue as a separate measure of market conversion. Combining the two without labels obscures precisely what investors need to understand.

The second-quarter accounts illustrate this distinction: GAAP revenue was $108.5 million, and price-protection income was separately reported at $17.6 million. Their sum should not be relabeled as customer sales. MP second-quarter financial statements.

Other government arrangements have different economics. The original package included $400 million of convertible preferred stock and a warrant, rather than an unrestricted grant to common shareholders. The separate 10X offtake establishes a contractual minimum defined EBITDA framework associated with the future facility; it is not a current company-wide profit guarantee. MP transaction filing, July 2025.

For common equity, that combination has both benefits and costs. The support can improve the ability to finance an industrial buildout and reduce exposure to a strategically vulnerable market. Preferred securities, warrants and borrowing also allocate rights to other capital providers. The appropriate question is how much reliable cash generation the completed system can produce after honoring those rights—not simply how large the announced public package appears.

11MP: liquidity provides runway while cash flow reveals the cost of expansion

At June 30, MP held $1.453 billion in cash, cash equivalents and short-term investments. Its convertible notes and samarium project loan had approximately $1.013 billion of combined principal, with equipment financing additional. These are different claims from the preferred securities. Customer concentration also remained significant: three unnamed customers represented 40%, 32% and 15% of second-quarter revenue, respectively. The filing does not authorize assigning those letters to particular named partners. MP Form 10-Q, June 30, 2026.

The balance sheet consequently offers substantial funding capacity without establishing that the expansion pays for itself. During the first half, operating cash flow was positive by just $4.9 million against $307.7 million of property, plant and equipment spending. The company’s defined free cash flow was approximately negative $302.8 million. Production and revenue can improve while the buildout draws on capital raised previously; this does not, by itself, establish a need for another financing. MP cash-flow statement and reconciliation, August 6, 2026.

Cash timing also matters. First-half operating cash flow included $93.3 million of price-protection receipts relating to the fourth quarter of 2025 and first quarter of 2026. Those receipts do not belong to the same period as all the protection income recognized in the accounts. Investors should reconcile the receivable, income and cash receipt rather than assume the reported operating cash flow represents a stable quarterly pattern. MP’s June 2026 cash-flow discussion.

The commercial concentration reinforces the need for careful ramp management. If one large customer’s qualification or ordering schedule slips, production and inventory could move differently even without a change in long-term demand. The government price mechanism can cushion eligible material pricing, but it does not automatically eliminate working-capital requirements or the cost of an unfinished downstream plant.

The next evidence to watch is therefore specific: commercially accepted finished magnets, repeat deliveries, stable separation output and a cash requirement consistent with the construction timetable. Greater output is most valuable when it moves through the next processing stage at an economic cost and ultimately becomes a collected customer invoice. That is how MP’s existing industrial advantage can become stronger cash generation per share; nameplate capacity alone cannot demonstrate the result.

12USA Rare Earth: owning Serra Verde changes the starting point

USA Rare Earth completed its combination with Serra Verde on September 3, 2026, and announced the closing the following day. That adds an operating Brazilian mine and processing business to the group. Pela Ema began production in January 2024; however, the closing announcement still described optimization and commissioning work, with a target of approximately 4,000 tonnes a year of total rare earth oxide equivalent at the end of 2026. A further expansion targets average production of 6,400 tonnes annually, with commissioning expected to start within twelve months. Those are different milestones from the acquisition itself. USA Rare Earth, September 4, 2026.

The distinction matters because Pela Ema produces mixed rare earth carbonate, or MREC. Reporting the rare earth content on a total oxide basis does not mean that the mine ships individual drums of separated dysprosium, terbium, neodymium and praseodymium oxide. Serra Verde’s August update specifically identified MREC as the product covered by its new offtake arrangement, with deliveries expected early in the fourth quarter. Subsequent separation remains an industrial step with its own equipment, reagents, recovery losses and quality requirements. Serra Verde, August 24, 2026.

Owning both ends of a future supply chain can improve coordination, but it does not make the intermediate chemistry disappear. The useful questions become how much usable material each batch contains, where it will be separated, how much survives the processing stages and which customer will accept the resulting product. A tonne of carbonate, a tonne of oxide equivalent and a tonne of finished magnet cannot be treated as interchangeable units in a production chart.

The transaction also changed the capital structure. The closing filing specifies $300 million in cash and 126,849,307 USAR shares as aggregate merger consideration. The surviving subsidiary assumed obligations under a DFC financing agreement with a maximum principal amount of $565 million. That commitment comprises separate tranches; it is not a statement that USAR received another $565 million of unrestricted cash at closing. DFC independently confirmed the financing agreement earlier in 2026. Closing 8-K, September 4, DFC financing confirmation.

For shareholders, the acquisition therefore has two simultaneous effects: it adds operating assets and it distributes ownership of the enlarged business across more shares. Assessing only the expanded resource base misses the second effect. Assessing only dilution misses the acquired operation. The test is whether the combination can turn feedstock into repeatable, appropriately priced deliveries while meeting the obligations attached to the purchase and its financing.

13USAR’s factories are at different stages of the production cycle

Stillwater’s Phase 1a magnet line was commissioned on March 26. At that time, USAR targeted 600 tonnes a year of run-rate capacity by the end of the fourth quarter of 2026. Commissioning demonstrates that equipment and process steps can work together; it does not prove that the line has subsequently operated at that annual rate or sold that quantity. Stillwater announcement, March 26, 2026.

The subsequent evidence is important. USAR’s September 15 risk update still reported no sales of neo magnets and stated that delays in metal feedstock from Less Common Metals had affected Stillwater production. The March milestone and the later disclosure can coexist: a commissioned production line can still face supply and commercial bottlenecks. SEC risk update, September 15, 2026.

Round Top must also remain separate from Pela Ema. The same filing describes the Texas project as being at the exploration stage, without construction or commissioning of the mine and related facilities. Its potential resource does not supply today’s factory simply because both sit under the same owner. Turning that resource into feedstock requires its own development, financing and execution.

Wheat Ridge addresses a different link. On July 14, the company reported commercial-grade dysprosium and NdPr oxide samples from machining scrap generated at Stillwater. These samples were intended for qualification at LCM, followed by conversion into metal and alloy. The facility is a demonstration plant used to validate processing routes and inform commercial engineering. The achievement demonstrates separation capability on that feedstock; it is not evidence that a full commercial separation plant is already supplying all downstream demand. Wheat Ridge update, July 14, 2026.

Recycling improves the industrial logic when recovered material can re-enter a process with reliable quality and acceptable cost. It also requires careful accounting. Scrap from manufacturing is not a new independent source of mined metal, and recovering it cannot create more rare earth content than entered the process. A useful operating disclosure would connect the quantity of scrap generated, the recoverable content, actual recovery and the amount returned to production. A broad recycling percentage without those denominators tells less.

Blacksburg is further from revenue. USAR broke ground on September 9 for a roughly $1.2 billion project targeting 6,400 tonnes a year of magnets and 5,000 tonnes of strip-cast metal and alloy. Commissioning is targeted to begin in 2028. Adding those figures to current output would erase the construction period and confuse two successive products in the same chain. Blacksburg groundbreaking, September 9, 2026.

The September 17 partnership with Pasqal and Riven Systems belongs in research, rather than current production. The proposed work combines automated chemistry experiments with models intended to identify better separation molecules, including comparisons between quantum and classical approaches. No measured commercial cost reduction was announced. It becomes economically relevant if validated chemistry improves a real process without sacrificing recovery, purity or reliability. Joint announcement, September 17, 2026.

14USAR’s June accounts describe the business before Serra Verde

For the quarter ended June 30, USAR reported $5.821 million of revenue and a $46.314 million operating loss. Cash and equivalents were $1.530 billion at quarter-end. These figures predate the September acquisition and cannot describe the combined group’s current balance sheet. They are nevertheless the latest reported quarterly baseline against which subsequent investment and financing must be reconciled. Second-quarter results, August 10, 2026.

The 10-Q explains that first-half revenue came entirely from LCM’s metal-making operations, rather than magnet manufacturing or mineral production. In the six months to June 30, operations used $75.324 million of cash; capital expenditure and equipment deposits consumed another $108.388 million. Financing provided $1.419 billion, primarily reflecting January’s equity placement after costs. Those are six-month flows, not quarterly spending rates. USAR 10-Q, filed August 10, 2026.

A large cash balance can coexist with an immature operating model because the balance may have been built through financing. The right bridge starts with cash at a stated date, adds subsequent actual receipts, subtracts acquisitions and project spending, and identifies commitments not yet paid. It should not add every announced loan ceiling, grant maximum and customer arrangement as though all were deposits in the same bank account.

Likewise, dividing June cash by one historical quarter of expenditure would not produce a dependable runway. The business perimeter changed; new facilities require construction spending; a mine and its financing entered the group. Even without an acquisition, equipment deposits can precede installation by several quarters. A smooth expenditure line would conceal the timing of large contractual payments rather than explain it.

The practical operating question is whether a growing proportion of spending starts producing repeat orders and cash collections. Qualification samples may be valuable even before meaningful revenue, but their value is evidence of progress rather than proof of profitability. Conversely, revenue alone is insufficient: sales that require expensive feedstock, rework or unusually high scrap rates can expand without improving cash generation. USAR needs both the commercial ramp and evidence that the manufacturing process works economically across successive batches.

15USAR’s government support must be separated into distinct arrangements

The Commerce Department’s June 3 final award comprises up to $277 million of federal incentives and up to $1.3 billion under a loan agreement. The government identifies Round Top, Stillwater and Blacksburg as supported projects. This is substantially more advanced than January’s nonbinding letter of intent, but a signed award still does not equal immediate payment. NIST and Commerce, June 3, 2026.

At June 30, USAR had received no direct-award disbursements and drawn no advances under the loan guarantee agreement. Shares and warrants had already been issued in connection with access to the arrangements. Funding is therefore accompanied by ownership costs and conditions, rather than being free capital. USAR 10-Q, government-grants note.

The September risk update adds two concrete financing checkpoints: substantial capital must be raised by March 31, 2027, and a revolving facility of up to $250 million established by June 30, 2027. It also says the bank facility supporting the Serra Verde buyer had not been documented, closed or funded. Its commitment was scheduled to expire on December 31, 2026 unless closed earlier, or sooner following specified offtake events. SEC update, September 15.

That buyer is a special-purpose vehicle, separate from USAR’s operating cash. The August 24 announcement described a $1.55 billion package combining a $750 million government investment commitment, a bank commitment for up to $500 million and a government forward purchase of at least $300 million over five years. The associated fifteen-year offtake covers Serra Verde’s Phase 1 production and includes price floors. These components have different counterparties, purposes and timing. SPV announcement, August 24, 2026.

A floor can protect a contract’s realized price against part of a market decline. It cannot manufacture missing tonnes or make every cost recoverable. The decisive evidence is delivery under the agreement, the financing of the buyer, settlement of invoices and compliance with the contract. Government backing strengthens the commercial framework; it does not replace that operating evidence. For USAR, watching those conversions is more informative than repeatedly adding the headline funding figures.

16Energy Fuels: uranium revenue and rare earth capability follow different clocks

Energy Fuels already operates an industrial processing platform at White Mesa in Utah. Its relevance to rare earths comes partly from experience handling the uranium and thorium associated with monazite, a rare earth-bearing mineral. The company describes existing infrastructure and licenses for managing those materials. That can remove an obstacle faced by a newcomer, but it does not turn uranium ore, monazite concentrate and separated rare earth oxides into the same product. Energy Fuels, rare earth processing overview.

In the second quarter of 2026, Energy Fuels sold 310,000 pounds of uranium concentrate at an average realized price of $80.48 a pound, producing approximately $25 million of uranium revenue. It produced 865,000 pounds of finished uranium during the quarter and approximately 1.7 million in the first half. Production, inventory and sales are therefore visibly different quantities even in the established uranium business. Second-quarter results, August 5, 2026.

The 10-Q reports no rare earth segment revenue in either the quarter or the first half of 2026. That fact does not erase earlier production achievements. It does mean that the latest reported sales cannot be used as evidence of a large recurring rare earth revenue stream. The balance sheet and revenue perimeter also precede August’s ASM acquisition. Energy Fuels 10-Q, filed August 5, 2026.

White Mesa has commercial NdPr separation capacity of up to approximately 1,000 tonnes annually. Heavy rare earths remain a different scale-up task. The July construction announcement targets capacity of approximately 20 tonnes of terbium oxide and 120 tonnes of dysprosium oxide per year, with those circuits expected by the end of 2027. Additional samarium, europium and gadolinium circuits are expected by the end of 2028. The stated design quantities depend on feedstock composition and recoveries. White Mesa expansion, July 29, 2026.

The distinction between output and capacity is particularly important for small-volume, high-value elements. A designed annual quantity assumes that the plant has suitable feedstock, operates as intended and recovers material to specification. Multiplying that quantity by an attractive spot quotation skips all three conditions, along with contract pricing and operating costs. It is a hypothetical gross-sales calculation, not company guidance or an estimate of available profit.

17UUUU’s qualification milestone is real; its scope is specific

On August 19, Energy Fuels said a Japanese permanent-magnet manufacturer had fully qualified its terbium oxide for commercial use, with no further qualification required by that customer. The manufacturer was not named in the announcement, so this remains an attributed issuer disclosure rather than a separately verified customer statement. It establishes a meaningful product-quality milestone; it does not disclose an order quantity, delivery schedule or contract price. Terbium qualification, August 19, 2026.

The earlier March production announcement explicitly concerned pilot-scale terbium made from U.S. monazite. The industrial sequence is therefore pilot production, customer qualification and commercial expansion, with each step solving a different problem. First terbium production, March 25, 2026.

Qualification asks whether the material works in a customer’s process. Scaling asks whether the supplier can repeatedly make enough of it. Commercial execution adds delivery reliability, price and payment. A qualified sample can reduce one important uncertainty without resolving the others. For a magnet manufacturer, consistency across lots can matter as much as the headline purity of the first sample: unplanned variation can interrupt a process designed around tightly controlled inputs.

The feedstock plan introduces another timetable. Energy Fuels reported a 12.7% interest in the Donald joint venture at June 30, with the ability to earn up to 49% through further investment, and rights to all monazite offtake. The July expansion plan assumed Donald production beginning in 2028, subject to a positive investment decision and financing. Ownership, offtake rights and material already delivered must remain separate categories. August 5 results, July 29 expansion assumptions.

That timing creates a useful monitoring framework. If downstream separation starts before sufficient feedstock is available, the plant needs another supply source or runs below capacity. If feedstock arrives first, inventory and working capital can build before sales. These are analytical scenarios, not forecasts. A credible integrated plan aligns mine development, processing readiness and customer demand; completing one element ahead of the others does not automatically optimize the whole system.

18ASM adds operating metal capacity; VAC remains a proposed acquisition

Energy Fuels completed its ASM acquisition on August 28, a status confirmed by ASM’s own implementation announcement. The Korean Metals Plant adds an operating metal and alloy business. Energy Fuels describes existing NdFeB alloy capacity of 1,300 tonnes annually and an expansion toward 3,600 tonnes, potentially commissioned as early as the end of 2026. The larger figure is a target, not the plant’s reported current output. ASM implementation, August 28, Energy Fuels closing announcement.

This is an acquisition of industrial capability rather than simply another mineral deposit. Converting oxide into metal, then producing an alloy with the required composition, occupies a separate manufacturing stage. Securing that capability can reduce dependence on an outside processor. It still requires compatible feedstock, operating discipline and a buyer for the output. The accompanying Dubbo resource should not be counted as present-day mine production merely because its owner has changed.

ASM’s annual report, posted on September 25, provides an unusually useful reality check: in the twelve months ended June 30, the Korean plant dispatched approximately 116 tonnes of NdFeB alloy and 33 tonnes of NdPr metal. These are shipments of different products, not nameplate capacity. The report also records first commercial sales of terbium and dysprosium metals to Neo Performance Materials. These pre-acquisition results belong to ASM’s reporting period; they do not retroactively become Energy Fuels’ second-quarter revenue. ASM annual report, page 3.

The completed ASM transaction issued 14,808,572 new Energy Fuels shares. The closing filing valued aggregate scheme consideration at approximately $243.4 million, including about $26.2 million in cash. These actual closing terms are more informative than recycling the original announcement’s share-price-dependent headline value. Energy Fuels 8-K, August 28, 2026.

VAC is different. The June agreement envisages $718 million in cash and 65.853 million new common shares, with possible additional preferred shares under specified pricing conditions. Closing was expected in early 2027, subject to approvals and other conditions. The intended purchase would add an established magnet business, including Sumter capacity; as of the cited updates it cannot be treated as a completed transfer of those factories into Energy Fuels. VAC transaction announcement, June 23, 2026.

Integration also changes how revenue should be read. Selling oxide between subsidiaries does not create a second external customer for the consolidated group. The economic benefit must emerge in the cost, reliability or margin of the final external sale. A longer corporate chain may retain more processing value, but it also inherits more equipment, inventory, employees and coordination risk. The important question is what the completed chain earns from outsiders after those costs, rather than how many internal transfers it can display.

19UUUU’s financing bridge: working capital is not an unrestricted construction budget

Energy Fuels reported approximately $996 million of working capital at June 30, including $58.4 million of cash and $878.3 million of current marketable securities. Working capital also reflects receivables, inventories and current liabilities; it is not a second cash account to add to those balances. The quarter’s consolidated net loss was $33.6 million. Second-quarter financial release, August 5, 2026.

During the first half, operations used $17.750 million of cash. Additions to equipment and mineral properties were $6.771 million and $17.287 million, respectively. Net cash from issuing common shares was $153.075 million. The $133.323 million investing outflow included purchases and maturities of securities, so treating all of it as factory construction would overstate that category. Energy Fuels 10-Q, cash-flow statement.

The balance sheet also carries financing obligations. Energy Fuels closed $700 million of convertible notes in October 2025, with a 0.75% coupon and 2031 maturity. The low coupon does not remove principal repayment or potential dilution; conversion and settlement follow the instrument’s terms. Convertible financing closing, October 3, 2025.

Separately, the June 18 announcement described a conditional commitment for up to $725 million of twenty-year senior-secured financing from the Office of Strategic Capital. Due diligence, definitive agreements, approvals and closing conditions remained outstanding in that announcement. It supports a potential funding pathway for White Mesa expansion and a U.S. metals facility; it is not reported cash already available to spend or an unconditional subsidy. Conditional government financing, June 18, 2026.

The next useful evidence therefore connects several accounts: actual funding draws, acquisition cash payments, plant expenditure, qualification-to-order conversion and external sales by product. Uranium provides a commercial business alongside the rare earth buildout, but its revenue should not be relabeled as rare earth success. Nor does an acquired alloy plant prove that every planned upstream expansion has been financed. Following those separate bridges allows readers to identify genuine industrial progress without confusing an integrated ambition with a fully synchronized production system.

20The comparison that survives a change in the share price

The four businesses should be compared along two axes at once: where the physical product is in the chain and how far the commercial evidence has progressed. Those axes prevent two opposite errors. An established producer should not receive credit for every downstream expansion as though it were already operating. A developer should not be described as having no progress merely because its final product has not reached customers. The right description identifies the completed milestone and the next unresolved dependency.

The table is a map for reading future disclosures, not a new set of company forecasts. For example, a quarter of rising group revenue can have several explanations: higher volume, a different product mix, an acquired business or improved prices. Those explanations have different implications for the original expansion thesis. A new subsidiary’s historical sales do not establish that another plant has passed customer qualification. Similarly, consolidated liquidity is not proof that every planned project has fully funded construction and working capital.

A practical reconciliation begins with the last reporting period before a transaction and the first after it. Note what entered the consolidation perimeter and for how many days. Separate transaction accounting from recurring operations, and identify whether management supplies comparable information for the acquired business. Then follow the actual product that motivated the investment thesis. If the thesis concerns heavy rare-earth separation, a rise in uranium sales is relevant to financing capacity but is not evidence of heavy rare-earth customer acceptance.

The same discipline applies to offtake. A contract can provide a future outlet and still leave price, recovery assumptions, acceptance or delivery timing unresolved. Its significance depends on what it obligates the buyer to take and the seller to deliver, not on the word “binding” alone. A volume allocation from a first project phase is different from an unconditional purchase of all later expansion output. This is why the counterparty’s account and the filed contractual conditions are useful alongside an issuer’s headline.

21What would close the information gap next

For CRML, the useful progression would connect the current ownership structure, the technical basis of the proposed refinery, the product qualification process and an identified financing plan. A new purity claim is most informative when accompanied by sample provenance, sustained throughput, recovery through the full route and independently reviewable technical detail. The question is not whether a headline sounds larger than the previous headline; it is which dependency has moved from an assumption to evidence.

For MP, a clear bridge between separated products, precursors and finished magnet sales would make the downstream transition easier to evaluate. Disclosed price-protection income should remain separate from customer revenue so that changes in commodity realization and changes in support mechanisms can be understood. Volumes, customer acceptance and production costs provide a stronger operating picture together than any one of those measures alone.

For USAR, the enlarged corporate perimeter makes integration and financing disclosures especially important. The reader needs to distinguish the Brazilian operation’s output from the Oklahoma plant’s commercial status, and completed funding from conditions still outstanding. For Energy Fuels, the key distinction runs through existing operations, recently acquired capacity and the timing of planned additions. A proposed acquisition can expand the strategic map before it expands the consolidated operating business.

An article written around these questions remains useful even when the four shares move sharply. It tells the reader what document to look for, what unit to check and what event would change the assessment. Political alignment can improve the context for investment, but material must still move through a qualified process to a paying customer. That is the evidence chain connecting a rare-earth story to a business a shareholder can actually evaluate.