A bag of tungsten concentrate was packed for shipment at a South Korean plant this month. That is a fact about a mine. It is not a fact about a stock price. On October 1, 2026, a note that began with Cantor Fitzgerald analyst Matthew O’Keefe made the rounds. He told clients that Almonty Industries has started shipping concentrate from Sangdong, is moving toward round-the-clock work, and has about 4.6 months of ore in a stockpile to feed the ramp. Phase II, he said, is already underway, with completion hoped for in 2027. Chief executive Lewis Black called it a defining moment. Tungsten mined and processed in an allied country, he wrote, is now a reality.
The Korea Times had the plainer version a few days earlier. On September 17, a Korean state inspector signed off on the processing plant and the crushers at Sangdong, in Yeongwol County. The site had been idle since 1993. The company said the first saleable bag, about a tonne, was packed and ready to ship. A permit is not a full mine. A first bag is not a year of steady sales. Both are further along than a drill plan.
This piece has one theme. Investors who want exposure to tungsten that is already being mined and bagged are looking at a producer, and Almonty is the producer in this story. Investors who want a higher-risk, higher-potential-reward claim on the same shortage are looking at an explorer that still has to prove its rock. Rackla Metals is one such explorer. Those are not two sizes of the same bet. One has customers, a plant, and a stockpile. The other has old holes and a plan to drill them again. The second belongs in a short section, because an unproven asset should not be asked to carry the same weight as a mine that is shipping.
Nothing here is advice to buy, sell, or hold Almonty, Rackla, or any other security. Cantor’s buy rating and its price target are Cantor’s. A junior can go to zero. A ramp can slip.
What “shipping” actually means
Read the verbs in order. Almonty has been running stockpiled ore through the plant since June 2026, according to later reports of the company’s own account. Until the September 17 certificates, that work was commissioning. It was not a commercial sale. The certificates changed the paperwork. The plant may now run, and the concentrate may now be sold at home or abroad. Black’s letter says the stockpile does more than feed the mill. It buys time, so the mining crews can turn to the next phase while the surface ore carries Phase I. Metso, the Finnish equipment firm, is on site to help tune the current plant. The same firm is being asked to supply mills for the expansion.
Phase II, in the company’s words, is officially a go. Underground development is moving. Orders are being placed. The target is completion in 2027, throughput of up to 1.2 million tonnes of ore a year, and potential output above 460,000 metric ton units of tungsten trioxide a year. A metric ton unit is 10 kilograms of contained oxide. Black said that scale would make Sangdong one of the largest tungsten mines outside China, and, in his phrase, the largest one producing today. Treat the last claim as a goal, not a present fact. A mine that has packed its first bags is not yet the largest producer in the world outside China. It is a mine trying to become one.
O’Keefe’s model puts numbers on that hope. Consolidated output, in his forecast, rises from 126,287 metric ton units in 2026 to 444,400 in 2027. All-in sustaining costs, in the same model, fall from $905 to $319 per metric ton unit. Cantor said these are forecasts, and they depend on the work going well. A cost that falls by about two-thirds while volume more than triples is a ramp model. Ramps are where mining stories go to be right or to be late. Power, labor, recoveries, and the grade of the ore that actually hits the mill can all move the $319. If they move it the wrong way, the stock does not get the year the model drew.
Why the West is paying attention
Tungsten is a hard, heavy metal. It keeps its strength when it is very hot. That is why it sits in cutting tools, drill bits, wear parts, and in parts of missiles, armor, and jet engines. You do not swap it out because a headline changed. China mines most of the world’s tungsten and, more important, refines a still larger share of it. In 2025 Beijing tightened exports of some tungsten products. Buyers in the United States, Europe, and Korea who need metal that does not come with a political string have been looking for other sources. A mine in an allied country that can fill a bag is the thing they have been describing in strategy papers. Sangdong is one of the few places where the paper has turned into a plant.
Two charts that traveled with the Cantor note are easy to misuse. They are from the International Energy Agency, drawn by Barclays. They show where copper, cobalt, lithium, nickel, graphite, and rare earths are mined, and where they are refined, in 2024 and in a 2030 case. They do not show tungsten. The pattern still matters. Mining is concentrated. Refining is more concentrated, and China dominates the refining bars. Tungsten fits that pattern even though it is not on those axes. A mine that only ships ore, or only ships a concentrate that must be refined in the same country the buyer is trying to avoid, has not broken the bottleneck. Almonty’s own map shows a “Sangdong downstream” box still marked under development. Concentrate is the step they have reached. A full Western refinery chain is a later step. Do not cash the later step today.
Defense analysts have been making the same point without the chart. Stifel’s Jonathan Siegmann has argued that the U.S. strategic stockpile of tungsten is badly run down, and that firms able to deliver supply from outside the covered countries stand to gain if Washington keeps trying to rebuild it. Bernstein’s Adrien Rabier has said European and Western rearmament is already under way, and that missiles, drones, jets, and tanks all need these materials. A stockpile order is not a purchase order with Almonty’s name on it. It is a reason the offtake talks happened. The offtake that is actually signed is more useful than the stockpile that is being discussed.
The contract is the business
A long-term offtake with Global Tungsten and Powders, part of Austria’s Plansee Group, has been reported as covering about 90 percent of Sangdong’s Phase I output. Later reporting said the deal was stretched to 21 years from first delivery, with a contracted volume of about 4.41 million metric ton units and a minimum of about 210,000 metric ton units a year after ramp-up. Phase I itself has been described as a plant built for about 640,000 tonnes of ore a year and roughly 2,300 tonnes of concentrate. Phase II has been described as a rough doubling of concentrate. Those figures come from company statements as carried by the trade press. They are not the same unit as Cantor’s 460,000 metric ton units. Concentrate tonnes and contained-oxide units answer different questions. An investor who adds them together will invent a mine that does not exist.
An offtake is a strength and a ceiling. It is a strength because a new mine with no buyer is a hope. It is a ceiling because more than 90 percent of Phase I is already spoken for, at a contract, not at whatever price a trader shouts on a tight day. If the tungsten price doubles again, Almonty may not receive the spot price on those tonnes. If the price falls, the contract is why the mine can keep running. People who buy the stock for a spot-price spike are buying a different asset from the one the contract describes. Phase II tonnes, if they are not yet fully sold, are the part of the story that still has price exposure. They are also the part that is not built.
Almonty is not only Sangdong. Its map shows Panasqueira in Portugal, acquired in 2016, in production. It shows Los Santos in Spain, on care and maintenance since the 2011 deal. It shows Valtreixal in Spain, still at pre-feasibility. It shows a molybdenum project at Sangdong, under development, acquired in the same 2015 package as the tungsten mine. The cash flow that can arrive in 2026 is Panasqueira plus the first Sangdong concentrate. The cash flow in the 2027 model is mostly the Korean ramp. One asset doing most of the future work is a concentration risk. A single mill, a single power line, or a single slow quarter in Gangwon Province can move the whole forecast. Portugal is the diversifier. It is not large enough to make the Korean ramp optional.
The target is an opinion
O’Keefe reiterated a buy rating and a 12-month target of $25.50. The note said that implied about 93 percent upside from the prior close it used. A target is a model plus a mood. The 93 percent is arithmetic on that model. It is not a return you are owed. Sell-side firms publish targets that miss. They also, at times, have banking relationships with the companies they cover. A reader who treats 93 percent as the investment case has skipped the only paragraph that matters, which is the one Cantor already wrote: the costs and the tonnes are contingent on execution.
The execution list is ordinary and long. The plant must recover the grade the mine sends it. The mine must send that grade on a clock, not in a lucky month. Metso’s mills for Phase II must arrive and work. The 4.6 months of stockpile must bridge the gap without becoming an excuse to delay the underground tonnes. Korean permits for the expansion must keep pace with the orders. The offtake partner must take the bags. The all-in cost must move toward $319, not stall near $905 while the share price trades as if $319 were certain. Any one of those can be fine and the year can still disappoint, because the model needs most of them at once.
There is a separate risk that has nothing to do with the mill. Tungsten prices have been high because exports from China tightened and because inventories were thin. High prices are why a closed mine from 1993 could be rebuilt. They are also why engineers look for substitutes in some tools, and why a new mine outside China, once it is really producing, adds supply. A successful Sangdong is good for Western buyers. It is not automatically good for the tungsten price. A producer with a low cost and a contract can live with a lower price. A producer that needed the scarcity to stay extreme cannot. Know which one the model assumes.
A junior is a different metal bet
Rackla Metals, ticker RAK in Canada, is the high-risk version of this shortage. Its tungsten ground is Lentung, once called Lened, in the Northwest Territories, near the closed Cantung mine. Union Carbide drilled it decades ago. Old reports show high grades. Rackla has said those figures are not fully verified, there is no current NI 43-101 resource, and it hopes to drill about 10,000 metres in 2026 and publish one by late 2026 or early 2027. A March note on this site that called the timing perfect. It cited about $9.7 million Canadian in cash. If new holes match the old grades, the gain for an early holder could be large, because the company is small. If they do not, the idea is over. Most explorers miss. Anyone who buys that risk should assume the stake can go to zero. Investos who picked up shares in early August for around 1 cents are not up over 100%.
How to tell the two apart
Ask one question. Can this company sell a bag of tungsten in the next twelve months without a new discovery. Almonty can, if the ramp holds, because the bag has already been packed and a buyer for most of Phase I has already signed. An explorer cannot. It can sell a story, and it can spend cash on drills. It cannot ship. Exposure to the metal, for an investor who wants the shortage to show up as revenue, means the company that invoices. Exposure to the hope that a shortage will make an old grade valuable means the company that still has to earn a resource. Mixing them because both say tungsten is how a careful idea becomes a careless portfolio.
The producer still has to be underwritten, not cheered. Read the 2026 tonnes against the 2027 tonnes and ask what has to be true in between. Read the cost drop the same way. Read the offtake and ask how much of the upside is already sold. Read Panasqueira as the mine that exists if Korea is slow. Read the downstream box on the map as work that is not done. Then, and only then, decide whether a mine in an allied country, at this stage of a ramp, is the tungsten exposure you meant. The analyst target is the last thing to read. It is a view. The bags, the contract, and the unbuilt mills are the business.
The close
Western tungsten supply has a new fact. A plant in South Korea that sat still since 1993 has a permit, a stockpile, a contractor, and a first bag of concentrate. Almonty says Phase II can take Sangdong toward 1.2 million tonnes a year and more than 460,000 metric ton units, with a 2027 finish. Cantor’s model turns that into a jump from about 126,000 metric ton units this year to about 444,000 next year, and a cost drop from $905 to $319. Those are forecasts. The offtake covers most of Phase I. The buy rating and the $25.50 target are one firm’s opinion, and the 93 percent is that opinion measured against one close.
Investors who want exposure to tungsten can look at a company that is already shipping it. Almonty is that company in this story, with the risks of a ramp, a concentrated asset, a sold-forward book, and a price that may not stay scarce. Investors who want higher risk and higher potential reward can look at an explorer that is not shipping, they're 'brownfields' exploring. Rackla is one of those. It is not a small Almonty. It is holes that have not yet been drilled to today’s standards. Only the producer is exposure to tungsten today. Keep the two in separate columns. The market will not do it for you.
Important information
This article is for information and education only. It is not an offer, a solicitation, or a recommendation to buy, sell, or hold Almonty Industries, Rackla Metals, or any other security. It does not consider any reader’s finances. Mining stocks can become worthless. Explorer shares are especially risky and often end at zero. Historical drill results are not a current mineral resource.
Figures on Sangdong’s first concentrate, the September 17, 2026 inspection certificates, and the idle period since 1993 are drawn from company statements as reported by the Korea Times and the trade press. Phase II targets, the 4.6-month stockpile, Metso’s role, and the 1.2 million tonne and 460,000 metric ton unit figures are from chief executive Lewis Black and from Cantor Fitzgerald analyst Matthew O’Keefe, as carried in an October 1, 2026 note. Output of 126,287 and 444,400 metric ton units, and costs of $905 and $319 per metric ton unit, are Cantor forecasts, which the firm said depend on execution. The buy rating, the $25.50 target, and the 93 percent upside figure are Cantor’s, measured from a prior close cited in that note. Offtake terms with Global Tungsten and Powders are as reported from company disclosures, including coverage of about 90 percent of Phase I and later extensions. They can change. The International Energy Agency charts refer to other minerals, not to tungsten. Rackla’s Lentung plans, including about 10,000 metres and a hoped-for resource estimate, are the company’s March 30, 2026 disclosure. The company said historical results are not fully verified. The cash figure is from a March 23, 2026 Canadian Mining Report note, and that note was promotional. This article is not a substitute for filings, a technical report, or a licensed adviser.