A missile line does not run on a press release. It runs on a pound of metal that a prime contractor has already agreed to take, in a form the Pentagon will actually put in a weapon. On October 1, 2026, Bloomberg Intelligence asked whether more than $40 billion in announced support for critical minerals will become that pound in time. The West is rearming. It wants missiles, bombs, drones, jets, submarines, and night-vision gear. China still sits on most of the refining. New mines take years. The note’s title is dry. The timing is not. Defense-critical mineral capital is moving downstream. The metal, in too many cases, is not.
This piece has one theme for investors. The $40 billion is not the supply. The supply is the next pound that is pure, qualified, and under contract. Almost none of the policy stack is that pound yet. China’s hold on refining is a through-2030 fact in the charts serious banks are using. A rearmament cycle that runs on the same clock will not wait for a ribbon-cutting. Buy the announcement and you own a hope. Track the qualified pound and you own the only object a defense buyer can spend.
Nothing here is a recommendation to buy or sell a miner, a magnet maker, or a stockpile. Analyst targets disagree with each other. Projects slip. A price that has already jumped can fall. The job is to separate a financing headline from a shipment.
What the forty billion is, and is not
Bloomberg’s analysts were plain. Execution, not government support alone, will decide whether U.S. policy turns into durable revenue and real readiness. Policy has moved past simple grants. The tools now include equity, price floors, loans, offtake deals, and stockpiles. Those tools are designed to keep capacity alive through a commodity bust. They do not dig a mine. They do not separate a heavy rare earth. They do not put a supplier on a jet’s approved list.
Read their map before you read a stock pitch. They scored materials by how critical a mission is, and by how available a defense-grade supply actually is. The scarce corner is the one that matters in a war. Dysprosium and terbium for magnets. Gallium for radar and electronic warfare. Antimony for munitions. Tungsten. Scandium for systems that are still mostly on paper. A more resilient band sits nearby. Neodymium-praseodymium magnets. Titanium for air and naval uses. Tantalum for electronics. Bubble size, in their chart, is how widely a material touches platforms and contractors. The positions are judgments, not GPS coordinates. The shape is still the lesson. Money is pouring toward a problem whose worst nodes are not the ones with a spare warehouse.
They named companies only as exposure, and then took the name back. MP Materials and ATALCO, they wrote, are the clearest near-term links to magnets and gallium. IperionX and Perpetua are targeted ways to own titanium and antimony stories. Then the sentence that should be on the first page of every pitch deck. Defense-grade output, a real customer, and contracted volume still have to follow the capacity that has been announced. A link is not a delivery. An investor who stops at the ticker has stopped one step too soon.
Five tools, none of them a missile
Each federal tool fixes a different market failure. Each one leaves the factory problem standing. Equity and preferred capital can pay for processing kit that private lenders will not finance. It keeps a domestic node from dying. It does not make the node good enough for a weapon. A price floor can stop a flood of cheap Chinese material from bankrupting a Western plant in a down year. It protects a building. It does not create a qualified product. A long offtake gives a new plant a buyer on paper, so the banks will lend. Paper demand is not a schedule a program office can march to. Loans and guarantees cut the cost of money and try to get a plant built before a cutoff. Construction still takes the years it takes. Stockpile finance tries to make firms hold emergency inventory they would rather not pay for. Inventory in the wrong chemical form is a pile, not a part.
That is the whole federal menu. It is serious. It is also a menu for keeping doors open. Readiness is what happens after the door is open, the assays are real, and a named prime has signed. Bloomberg’s own ladder says the same thing in four steps. First commercial output only proves the plant can run. Purity and yield are not permission to bolt the product onto an F-35. A qualification award, with a named contractor or a named program, is the first step into funded procurement. Contracted shipments, with volume, price, and a date, are when a story becomes revenue a defense investor can model. A second source is what reduces the chance that one fire or one export ban ends the line. Most press releases live on step one. The stock charts often price step four.
The jet already knows the weight
The analysts noted that an F-35 needs more than 900 pounds of rare-earth material. One aircraft. Not a fleet. Not a war. Magnets sit in actuators, generators, and motors. Missiles use them in control and guidance. Drones use them because a cheap airframe still needs a motor. Satellites use them in reaction wheels and power systems. Submarines use them in electric motors that are now part of the shipbuilding clock. Lose the magnet and you do not lose a slogan. You lose production, a munitions ramp, or a hull schedule. The platform table is a list of ways a missing pound becomes a late weapon.
Heavy rare earths are the sharp end. Dysprosium and terbium are what let a magnet keep its strength when it gets hot. Fighter and missile magnets get hot. Separation of those heavies, yields in the plant, and the slow work of customer qualification are the bottleneck under the bottleneck. Mining the ore in a friendly country does not finish the job. Bloomberg said this in one line, and it is the line investors skip. Refining and downstream production are still the critical part. A mine without a separator is a quarry with a flag on it.
China’s clock runs through 2030
Christian Keller, Barclays’ global head of economics research, has said the quiet part without a costume. Breaking China’s quasi-monopoly in these materials is unlikely to be a story this decade. Mining and refining concentration, in the charts he and the International Energy Agency are pointing at, persists through 2030. Look at the mining chart and you see a crowded room. Copper, cobalt, lithium, nickel, graphite, rare earths. Australia, Indonesia, the Congo, Chile, China, and a handful of others. Concentrated, yes. Not a single flag over every pit.
Look at the refining chart and the room empties. China’s share of processing is the red bar that does not go away by 2030. Copper, cobalt, lithium, nickel, graphite, rare earths. Other countries appear as slivers. Russia, Finland, Chile, Indonesia, Malaysia. The rest of the world is a remainder. This is the picture a rearmament plan has to live inside. You can announce a domestic mine in 2026. The chemical plant that turns the mine into a defense part is still, on these charts, mostly in China at the end of the decade. An investor who buys “ex-China” on a slide and does not ask where the concentrate is refined has bought a passport, not a supply chain.
Tungsten is the cautionary receipt
Stifel’s aerospace and defense analyst Jonathan Siegmann told clients last week that investors want to own the bottlenecks. He meant the producers that can deliver now, not the ones that can deliver in a keynote. The chart he put under that line is a national embarrassment with an axis. The U.S. National Defense Stockpile of tungsten fell from about 37,000 tonnes to about 1,500 tonnes. The title is the argument. A peace dividend paid in tungsten. A 25-year drawdown that will not be reversed quickly. In the early 1990s the Pentagon decided it held too much and sold. Ores, powders, carbide. The sales ran for years. Then the pile sat. Then, late, someone noticed that a war uses tungsten and the cupboard is a few months of unprocessed ore.
China, on the figures Stifel has used, held about 80 percent of mine supply in 2025 and about 85 percent of ammonium paratungstate refining. Export controls helped drive tungsten prices up on the order of 775 percent from the start of 2025. That jump is a fact about the past. It is not a promise about the next bid. The U.S. government has said it wants to buy again. It also has to buy from a domestic base that, in Siegmann’s telling, largely does not exist yet. You cannot restock a 25-year hole with a purchase order and a wish. The hole is the reason the metal is strategic. The hole is also why a headline about “the West’s tungsten champion” deserves a calendar check.
The miner in that headline is Almonty. The banks do not agree about it, which is the point ZeroHedge’s last line smoothed over. Stifel initiated with a Buy and a $25 target. Its case is that Sangdong in South Korea has begun Phase I production, Panasqueira in Portugal is expanding, and the company could lead Western tungsten output by the end of 2028. Goldman Sachs initiated at Neutral with a $13 target. The published reason for the caution is that a lot of hope is already in the price, and that the ramp may be slower, and the price spike less permanent, than the bull case needs. Jefferies has covered the name. Coverage is not the same word as a buy. A $12 gap between two famous targets is not a chorus. It is a fight about timing.
The operations are real and they are early. Almonty has said tungsten was pulled at Sangdong for the first time since 1993. The crushing and processing plant received its operating certificates. By late September the Korea unit had a first tonne of saleable concentrate ready to export. A tonne is a start. It is not a stockpile. It is not 37,000 tonnes put back in a U.S. warehouse. It is not defense-grade metal in every form a shell factory uses. Stifel’s “largest in the West by 2028” is a forecast with a date on it. Forecasts move. Certificates can be followed by a slow ramp. An investor who needs the bottleneck today is allowed to like the mine and still refuse the verb “is.” The verb is “might, if the ramp holds, in two years.” Might is not a shipment.
A vault is not a magazine
Project Vault is the other place investors are being invited to relax. It was announced in February 2026. The Export-Import Bank approved up to $10 billion in financing. Private capital was meant to sit alongside it. The aim is a commercial reserve of critical raw materials for U.S. manufacturers, a cousin of the National Defense Stockpile, not a replacement. In September, Glencore and Mercuria became the first two trading houses inside VaultCo, the private firm set up to run it. Each committed $500 million. That is $1 billion of intent to source and stockpile. It is not $1 billion of metal already in a shed. CSIS, which has followed the program, said the open question was always how financing authority becomes an operational reserve. September’s news is a start on that question. It is not the answer.
Bloomberg’s comparison table is the adult version. In a Chinese export cutoff, Vault members may draw and later refill. The defense stockpile may release what it is allowed to release. The question that matters is whether the inventory exists in a defense-qualified form. The bottom line on the table is partial. Feedstock, maybe. Qualified form, limited. In a wartime surge, nobody has published a sizing that matches the consumption. In a single-supplier failure, Vault can be fast for its members and the national stockpile can be slow, and form still governs. In a price bust, inventory finance can keep a firm alive. It does not, by itself, keep a conversion plant alive unless domestic-content rules or a price floor do the rest. A commercial buffer is a good idea. It is a bad substitute for the four-step ladder. A drone motor cannot be wound with a financing commitment.
The supercycle will not pause for commissioning
Adrien Rabier, Bernstein’s analyst for European aerospace and defense, has put a clock on the allied side of this. Europe’s defense spending supercycle is already ramping. He has it running through 2030. That is the same year Barclays says the refining concentration is still with us. Two clocks, one collision. Governments will order ships, shells, and aircraft on a political timetable. Mines and separator plants arrive on a geological and permitting timetable. The gap between those timetables is where programs slip and where investors overpay for stories that are true too late.
The collision is not a reason to ignore the sector. It is a reason to change the question. Do not ask whether the West has announced enough money. Bloomberg already framed the number as more than $40 billion, and the number will be cited in every slide. Ask which pound has independent purity data. Ask which supplier has a named prime. Ask which contract has a volume and a date. Ask which material has a second source that is not in the same political blast radius. Those four questions are boring. They are also the only ones that match how a defense buyer spends. Everything else is a narrative about a future pound.
Tungsten shows the habit in one metal. The stockpile was sold down over a generation. The price then screamed. A Korean mine that slept since 1993 shipped, or stood ready to ship, a first tonne. One bank sees a Western leader by 2028. Another sees a stock that already reflects the dream. Neither bank can put a shell on a pallet this quarter. Rare earths show the same habit in a heavier form. The magnet metals that sit in the scarce corner of Bloomberg’s chart are not solved by a light-rare-earth headline. Gallium, antimony, and the rest of that corner are not solved by a vault that holds the wrong chemistry. The pattern is the investment. The pattern is that capital has moved faster than qualification.
What to watch, and what to ignore
Ignore the round number when it is used as a finish line. Forty billion dollars of announced support is a measure of political effort. Effort is not oxide. Ignore a mine’s “ex-China” label if the refining step is still a Chinese plant. Ignore a price chart that starts in a panic and is drawn as a destiny. The tungsten move since the start of 2025 is a warning about export controls. It is also a number that can reverse if controls ease, if scrap rises, or if a ramp arrives faster than the bulls need it to.
Watch four pieces of paper. A lab result that is not the company’s own slide. A qualification letter with a contractor’s name on it. A contract with tonnes, a price, and a quarter. A second source that could ship if the first source stops. Watch Project Vault for a different kind of proof. Not another commitment. A published list of what is actually in inventory, in what form, and who is allowed to draw it in a cutoff. Watch the defense stockpile for tungsten in particular. Rebuilding from about 1,500 tonnes toward anything like the old pile will be visible, or it will not. Absence of purchases is also data.
Watch the disagreement. When Stifel says $25 and Goldman says $13 on the same tungsten miner in the same week, the spread is the analysis. One house is paying for 2028 leadership. The other is refusing to pay twice for a ramp that has just produced a tonne. You do not have to pick a side to learn the lesson. The lesson is that “the West’s largest supplier” is a sentence about a year that has not happened. Sentences about years that have not happened are how this trade goes wrong.
The close
The West has decided, late, that it does not like depending on China for the materials inside its weapons. It has put a large number on that decision. Bloomberg’s number is more than $40 billion in announced support. EXIM’s number for one vault is up to $10 billion in financing authority, plus a first $1 billion of trader commitments that are not yet a pile of metal. The Pentagon’s tungsten cupboard is the ghost of a larger pile. A Korean mine has a first tonne and a forecast. Europe’s arms orders run toward 2030. China’s refining share, on the charts Barclays is using, runs toward 2030 as well.
Those facts can all be true and the investor can still buy the wrong thing. The wrong thing is the announcement. The right object is smaller, duller, and harder to find. It is a pound that has been assayed by someone who does not work for the seller, accepted by a named program, and scheduled in a contract. Until that pound exists, the rearmament is a demand story and the minerals policy is a supply story, and they do not meet. The theme is the gap. Capital has moved. Qualification has not. China still owns the middle of the chain. A stockpile sold off over twenty-five years will not be bought back in a news cycle. Own the bottleneck only if it can deliver. If it can only announce, you do not own the bottleneck. You own the wait.
A note on sources
The $40 billion figure, the scarce-versus-resilient map, the five federal tools, the magnet platform table, the four investor milestones, the F-35 rare-earth weight, and the company examples are Bloomberg Intelligence’s, from the note described in reporting on October 1, 2026. They are analyst judgments, not a government census. Keller’s through-2030 view and the mining and refining charts are Barclays’ reading of International Energy Agency data. Siegmann’s stockpile chart, the drop from about 37,000 tonnes of tungsten to about 1,500, and the “own the bottlenecks” line are Stifel’s. The 80 percent mine share, 85 percent refining share, and roughly 775 percent price rise are figures used in Stifel’s Almonty initiation, as reported, not a price forecast. Stifel’s Buy and $25 target, and the end-2028 Western-leadership case, are that initiation. Goldman’s Neutral and $13 target are Goldman’s, and they are not a buy. Jefferies is among the firms that cover the stock. Coverage is not a rating. Sangdong’s first extraction since 1993, the plant certificates, and the first saleable tonne are company reports from late September 2026, not a completed ramp. Project Vault’s February 2026 launch, EXIM’s financing of up to $10 billion, and the September commitments by Glencore and Mercuria of $500 million each are from the bank, the firms, and CSIS. Those commitments are not proof of delivered inventory. Rabier’s European timeline is Bernstein’s. This is not advice to buy or sell any security. Mines miss dates. Export rules change. Targets are opinions.