Vanadium's Long-Term Price Is a Barbell, Not a Cycle

A price that halved while the story improved

Vanadium pentoxide averaged $12.17 a pound in 2019 and $9.25 in both 2022 and 2023 on the Chinese benchmark that USGS carries in its mineral commodity tables. By September 2025 the same benchmark was quoted at $4.86 a pound — 47.5% below its 2022 level and 60.1% below 2019 — while the long-duration energy storage industry that is supposed to transform vanadium demand kept announcing projects, and vanadium redox flow batteries kept winning orders against lithium in the duration segment where their economics make sense.

That divergence is not a market that has failed to notice the story. It is a market with two prices in it, and only one of them has anything to do with batteries.

Ninety-four percent of this metal is a steel order

In the United States, metallurgical use — overwhelmingly as an alloying agent for iron and steel — accounted for about 94% of reported vanadium consumption in 2023, on USGS’s own account. It is the only market that publishes the split cleanly, but the metallurgy it describes is not American: vanadium’s largest single use everywhere is as a micro-alloy that raises the strength and fatigue resistance of steel, at concentrations measured in grams per tonne.

That has a specific consequence for anyone forecasting price. Vanadium demand is not a function of industrial sentiment or of the battery narrative. It is a function of two numbers multiplied together: how many tonnes of steel the world makes, and how many grams of vanadium per tonne the specification requires. The first is cyclical and currently soft, particularly in China, where construction-linked rebar demand has been the dominant volume. The second moves only when a standard or a design changes.

The rest of the demand stack — catalysts for maleic anhydride and sulphuric acid, aerospace titanium alloys, chemicals — is small in tonnage and slow-moving. It is also, in the aerospace case, close to irreplaceable: USGS notes that while manganese, molybdenum, niobium, titanium and tungsten can substitute for vanadium to some degree in steel, no acceptable substitute exists for vanadium in aerospace titanium alloys.

Two countries produce most of it, and neither producer decides to

World mine production was about 100,000 tonnes of vanadium content in 2023, rounded. China produced 68,000 tonnes of that — 68% — Russia 20,000, South Africa 9,100 and Brazil 6,400. The United States produced none from primary ore and met 58% of apparent consumption from imports.

Supply concentration of that order usually implies pricing power. Here it implies the opposite, because of how the metal is made. USGS is explicit that vanadium “is typically recovered as a byproduct or coproduct”, and that demonstrated world resources are therefore not fully indicative of available supplies. The Chinese tonnage is overwhelmingly a co-product of vanadium-bearing iron ore processed for steel: vanadium comes out of the slag because the steel is being made anyway, not because somebody decided the vanadium price justified the mine.

So the supply curve has an unusual shape. When vanadium prices fall, co-product output does not fall with them, because the decision that produces it was a steel decision. Output only declines when steel output declines, or when a producer is pushed to curtail by cash losses rather than by price signals — which is what has been happening, first in the West since 2023 and then in China through 2025.

The trough, in one company’s accounts

The clearest way to see what a sub-$5 pound does to the primary end of the industry is to read Largo Inc.’s 2025 results. The company operates the Maracás Menchen mine in Brazil and is one of the largest primary vanadium producers outside China and Russia.

In 2025 it produced 9,150 tonnes of V2O5 equivalent, within its guidance range of 9,000 to 11,000 tonnes, at a global recovery rate of 80.1%. It sold 8,686 tonnes. Because Largo reports in V2O5 equivalent — the traded unit, which is 56% vanadium by mass — that production is about 5,124 tonnes of vanadium content, or roughly 5% of world mine supply. It is a meaningful asset, not a marginal one.

The financial outcome was a net loss before tax of $51.7 million, which works out at roughly $5,949 of pre-tax loss per tonne sold. Against a benchmark of $4.86 a pound, a tonne of V2O5 was worth about $10,714. Even allowing for non-cash items, impairment and financing costs in that loss, the direction is unmistakable: a tonne sold at roughly $10,700 carried a pre-tax loss of roughly $5,900. The company ended the year with $9.7 million of cash against $107.1 million of debt, and in October 2025 secured a term sheet with five Brazilian lenders representing $84.2 million of that debt to defer principal repayments to March 2026, with an automatic rollover to September 2026 conditional on raising at least $22 million.

This is what price-inelastic supply looks like from the inside. A primary mine cannot switch its output off and on: the orebody is fixed, the cost base is largely fixed, and the debt service is entirely fixed. So it keeps producing into the trough, which is precisely the behaviour that keeps the trough deep. The co-product producer keeps producing because steel is the decision. The primary producer keeps producing because the balance sheet is the decision. Neither responds to the vanadium price, and the market clears at a level that hurts the second group until enough of them stop.

The demand that could break the link

Vanadium redox flow batteries are the only demand segment with a plausible claim to reprice this metal, because the electrolyte is vanadium. In a flow battery the energy capacity is set by the volume and concentration of vanadium in the tanks, not by the stack, which means the marginal cost of an extra hour of duration is a chemical purchase. That is the structural reason flow batteries compete in long-duration storage and lose in short-duration storage: their cost curve is a vanadium curve.

The demand shift is real but has been slower than the projections. USGS’s commodity summary cites third-party projections of VRFB taking roughly 17% of vanadium consumption by 2033, against about 3% in 2021. It also makes the observation that matters more than the percentage: the new supply of high-purity vanadium pentoxide needed to serve that market “was expected to come from either existing producers or from projects which are, for the vast majority, still in their very early phases of development.”

Read that sentence twice. The demand growth is forecast with a date attached. The supply response is not. Greenfield vanadium projects have permitting, financing and offtake timelines measured in years, and the current price environment is the worst possible moment to finance one — which is exactly how commodity markets produce the shortages that later justify the projects nobody financed.

Two prices, not one

Here is the part of the vanadium market that most analysis flattens. Metallurgical vanadium and battery vanadium are not the same product.

Steelmakers buy ferrovanadium or vanadium nitride — an alloy or a nitride, produced from slag or from V2O5, with a specification written around steelmaking. A battery electrolyte needs high-purity vanadium in solution, with tight limits on the other metals that would degrade a circulating electrolyte over a twenty-year asset life. Different specification, different qualification, different logistics, and a different set of qualified suppliers.

That matters because the two products have different marginal cost structures. Metallurgical vanadium from co-product slag has a cost base close to the cost of the steel operation that produced it. Battery-grade vanadium requires purification and, in most Western jurisdictions, a supply chain that can demonstrate provenance. The price premium a battery customer pays is not a market inefficiency; it is the cost of that purification plus the scarcity of qualified supply.

Which is why a single “vanadium price” is a poor contracting instrument. The number that matters depends on which product you are buying, and the spread between them widens exactly when battery demand accelerates.

The barbell, and what to do with it

Put the two halves together and the long-term price structure stops looking like a cycle and starts looking like a barbell.

The floor is the full cost of the marginal primary producer able to deliver battery-usable purity at scale. That producer is not Chinese slag; it is a mine like Maracás or a refiner built for the purpose, and its cost does not fall because the spot price falls. The ceiling, in the near term, is set by co-product economics that are close to zero, which is why the market can sit below the primary cost base for years at a time without a shortage appearing.

Between those two, the price overshoots in both directions. That is the environment we are in: a spot market at $4.86 a pound, a primary producer losing $5,900 a tonne to supply it, and a battery industry whose growth case depends on a purity grade the spot market does not price.

The rational response from anyone who needs vanadium for twenty years is therefore not to buy at spot. It is to move off spot in one of three ways. Index with a floor: contract on a published benchmark but with a price floor at the producer’s cost of production, so the offtaker absorbs the downside only to the point where supply survives. Tolling: the offtaker owns the metal, the producer rents the conversion, and neither party is exposed to the flat price. Electrolyte leasing: the vanadium is leased rather than sold, and returned at end of life, which converts a commodity exposure into a service charge and removes the metal’s volatile capital cost from the battery’s levelised cost.

The industry has begun to move this way. Largo holds a 50% interest in Storion Energy, a venture with Stryten Energy focused on US electrolyte production for utility-scale flow batteries, and in June 2025 Storion signed a strategic supply agreement with TerraFlow Energy targeting initial energisation in early 2027. That is a producer choosing to sell a converted product with a contract attached rather than a commodity priced at a benchmark it cannot influence.

The company is also diversifying the same orebody. In September 2025 it began installing additional flotation circuits to raise ilmenite concentrate capacity from 42,000 to 115,000 tonnes a year, with the circuit resuming operation in November. A titanium feedstock co-product inside a vanadium mine is not a hedge in the financial sense, but it is a second revenue line on fixed infrastructure — which is the only kind of cost relief available to a primary producer in a trough.

What to watch

• Chinese steel output. It is the single largest variable in vanadium demand and the reason the co-product supply base will not contract on its own. Rebar and construction activity matter more than battery announcements.
• Purity supply, not tonnage. The question for battery demand is who can deliver high-purity V2O5 at scale and on contract. Watch qualification announcements, not resource statements.
• The Western primary balance sheets. Largo’s debt deferral runs to March 2026 with a rollover to September conditional on raising $22 million. How that resolves is the market’s cleanest test of whether primary supply survives the trough intact.
• Project financing. A greenfield vanadium project reaching a final investment decision at $4.86 is the strongest possible demand signal, because it would mean somebody can see a contracted price above the cost of building new supply.
• The spread. If a battery-grade premium over metallurgical V2O5 becomes visible and persistent, the barbell is real and the single-benchmark era is over.

The long-term price is not the spot price extrapolated

Vanadium is a metal where 94% of demand is an invisible micro-alloy in steel and 100% of the recent narrative is about batteries, and neither of those facts sets the long-term price on its own. The steel side sets the ceiling, because co-product supply is almost free. The battery side sets the floor, because the supply that can meet a battery specification is not free at all.

Which means the useful long-term number is not a forecast of the spot price. It is the cost of producing a high-purity tonne from a primary asset, plus a scarcity premium for the small number of suppliers who can qualify — and any contract structure that does not reach that number is not a long-term price at all. It is a speculative position with a delivery obligation attached.