Vanadium Pentoxide Price Trend: September 2026 Update

Updated 7 September 2026 · Assessments checked against CBCIE Metal, FerroAlloyNet, ScrapMonster, USGS MCS 2026 and Investing News Network on the same date · 15 questions answered at the foot of the page.

Where the number sits this week

Published vanadium prices are assessments, not settlements. There is no exchange contract, so every figure below is one reporter's view of where business could be done on a given day, on a given basis. That distinction matters more here than in most metals, because the bases are genuinely far apart.

The two-day move in ferrovanadium is worth pausing on. On 2 September, FeV50 was quoted at RMB 81,000–83,000/t. By 4 September it was RMB 83,000–84,000/t, with FeV80 up RMB 2,400–3,200 over the same window. That is roughly a 2.5% move in forty-eight hours, on the back of mill restocking rather than any supply event we can identify. If you are pricing a contract off a weekly snapshot, a move like that decides whether you make money on the lot.

The 2026 path, month by month

Read plainly: up about 14% from January to April, then down about 10.5% from the April high to July. The whole 2026 range so far is USD 5.30 to 5.81 per pound, which is a narrow band by vanadium's historical standards - this is a metal that traded above USD 9/lb in early 2022.

There is a curiosity in that table that we would flag rather than smooth over. The May and June assessments both read USD 5.80/lb while the tonne figures differ by nearly USD 600. Two series that are supposed to be the same price expressed two ways do not reconcile at those two points. Our working assumption is that the per-pound series is rounded or smoothed more coarsely than the per-tonne one. Treat them as two separate indicators, not as one number converted twice.

Three ways to misread these numbers

1. Pounds of oxide versus pounds of vanadium

V₂O₅ is 56.01% vanadium by mass - 181.88 g/mol, of which 101.88 g is the metal. A price of USD 5.00/lb of oxide is therefore about USD 8.93/lb of contained vanadium. Some contracts, particularly in the US ferroalloy trade, are struck on contained metal. The difference is a factor of 1.785, which is not a rounding error. Write the convention into the contract in words, not in a tick-box.

2. The implied exchange rate hiding in a dual quote

Here is a check we run on every dual-currency assessment. Take RMB 71,500/t and USD 4.78/lb. Converting: 4.78 × 2,204.623 = USD 10,538/t. Then 71,500 ÷ 10,538 = 6.78 CNY/USD. That is the exchange rate the source is implicitly using. If your own treasury converts at 7.10 or 7.15, then the two figures on that page are about 5% apart from each other before anyone has moved a tonne of material.

We are not criticising the source; reporters use whatever rate convention their subscribers expect. We are saying: do not take the RMB number and the USD number from the same page and treat them as interchangeable. Pick one, use your own rate, and say so in the contract.

3. Cash, acceptance, and tax

Chinese domestic quotations carry a basis string that changes the economics materially: "ex-works with tax by acceptance" means the price includes VAT and settles with a bank acceptance bill, not cash on delivery. A 90-day acceptance at a 4% annual discount rate costs roughly 1% of invoice value. On a RMB 71,500/t lot that is about RMB 715/t of financing embedded in what looks like the same number as a cash price. When a Chinese and an international quote differ by less than that, the difference may be the payment instrument rather than the molecule.

What the alloys say about the oxide

This is where the interesting arithmetic lives, and it is the kind of calculation that generic market commentary does not do.

From FeV80 back to flake. One tonne of FeV80 contains 800 kg of vanadium. Flake at 98% V₂O₅ carries 0.98 × 0.5601 = 54.9% vanadium. At a 95% metallurgical recovery, the flake needed per tonne of FeV80 is 800 ÷ 0.549 ÷ 0.95 = about 1,534 kg. At RMB 71,500/t, that is RMB 109,700 of oxide inside a tonne of FeV80.

Against a transacted FeV80 price of RMB 129,600–131,200/t on 4 September, that leaves roughly RMB 20,000–21,500 to cover reductant, electricity, labour, refractory, yield loss on the remaining 5%, and margin. It is not a fat spread. It does, however, tell you something useful: if oxide rises by RMB 5,000/t and FeV does not move, the smelter's whole contribution is gone. Smelters know this, which is why oxide rallies in this market tend to be short and contested.

The VN loss is the more revealing signal. FerroAlloyNet's weekly profit table for 31 August to 4 September prices VN at RMB 107,000/t against a computed cost of RMB 109,474/t, using retail flake at RMB 70,250/t. A loss of RMB 2,474/t. The mid-chain is currently paying for the oxide's stability with its own margin.

Two readings are possible and we hold both lightly. Either the alloy price is about to follow the oxide up, which would confirm the tightness that CRU and others expect later this year. Or the alloy price cannot be raised because rebar demand will not take it, in which case VN producers cut run rates and the oxide loses a customer. The 2025 numbers point the second way: VN consumption fell 3.8% to 36,690 t last year even as FeV50-equivalent consumption rose 10.4% to about 39,985 t.

Why higher prices are not calling forth much supply

The textbook answer to a price rise is more production. Vanadium does not work that way, and the 2026 evidence is unusually clear.

Stone coal is effectively dead. Output fell to 7,600 t in 2025 from 10,200 t in 2024, and the route is loss-making across the board. As of 2026, one large producer was still running at 100–120 t/month of ammonium metavanadate on a V₂O₅-equivalent basis; a Shaanxi plant with 300–350 t/month of capacity has been shut since early 2026 over safety issues. The kicker, per The Metalnomist: even when flake reached about RMB 110,000/t in 2023, stone coal output only reached roughly 11,000 t. That is what a real price signal bought. It is not much.

The slag route is feed-constrained, not capacity-constrained. China's total vanadium capacity was 277,600 t in 2025 against actual output of 163,900 t, so there is nominally 70% of headroom sitting idle. It does not get used because the binding input is vanadium-titanium magnetite, not roasting capacity. Panzhihua-area ore supply is reported as tight enough to cut 4,500–5,000 t of V₂O₅ equivalent in 2026; some Sichuan and Yunnan producers are having to source from Chengde or import instead. Against that, Qinhuangdao Baigong commissioned a 10,000 t/yr line in early 2026 with guidance of about 5,000 t for the year. Add one, subtract one.

Secondary supply is a by-product of molybdenum. Recovery from spent catalysts and residues rose to 15,100 t in 2025, up 1,900 t, including about 6,700 t from alumina by-product recovery. The reason it held up despite weak vanadium prices is that the same feeds carry molybdenum and tungsten, and those prices were strong. Vanadium came along for the ride. Nobody builds a recovery circuit for vanadium alone.

Demand: the storage story is real and currently overrated in the price

China's VRFB installations for 2026 are preliminarily estimated at 4–5 GWh, requiring on the industry's own numbers some 32,000–40,000 t of V₂O₅ equivalent. Vanitec has projected cumulative Chinese capacity above 8 GWh this year, with 35,000–40,000 t of oxide consumed. On any reading that is the largest single source of incremental demand, and it is why the storage share of Chinese vanadium use went from 4% in 2021 to 20% in 2025.

And yet the oxide sits below last year's annual average. A view we hold and will state plainly: the storage bid is being absorbed by inventory building at electrolyte plants rather than by spot purchases of flake. Chinese electrolyte project capacity was reported above 4.5 million cubic metres a year as of May 2026, across Rongke Power, Hunan Yinfeng, Zhejiang Xingchen and Jianlong. Plants at that scale fill tanks once, not continuously, and they fill them ahead of projects that are still under construction. The demand is real; its timing in the spot market is lumpy and easy to misread as strength that has already happened.

We would also give the sceptics their hearing. Amy Bennett of Fastmarkets, speaking at the Fastmarkets Global Battery and Critical Materials Conference in Las Vegas and quoted by Investing News Network in August 2026, argued that vanadium is "starting to lose out in this conversation" outside China as the storage sector shifts decisively to LFP chemistry, and that flow batteries are "just not hearing people talk about them so much anymore". Anyone modelling a lithium-style demand curve for vanadium should sit with that quote for a while.

What we are not going to do

We do not publish a Q4 2026 price target. Not because we have no view, but because a specific number published on a supplier's website is worth less to you than the method, and because the honest range is wide: the gap between the deficit case and another year of flat Chinese steel demand is several dollars a pound.

We will say where the arguments sit. For a firmer market: steel demand stabilising (worldsteel's April 2026 outlook points to 0.3–1.3% global growth in 2026), Chinese vanadium intensity rising to 51 g of vanadium per tonne of crude steel in 2025 from 48 g in 2024, rebar licensing rules taking effect 1 April 2026 that should favour vanadium-nitrogen micro-alloyed product, and a supply side that is feed-constrained. Against it: Chinese rebar output fell 4.5% to 186.3 Mt in 2025, the USGS records 2025 world mine production down at 110,000 t from 118,000 t in 2024 for reasons of policy rather than geology, and Western storage demand for flow batteries has gone quiet.

How buyers are actually managing the exposure

Since there is no listed hedge, exposure is managed structurally:

• Short validity windows. Export indications valid three to seven days; domestic tender prices valid for the tender week. Anything longer is a bet you did not mean to take.
• Formula pricing. Link to a named published assessment, state the conversion convention, and state what happens if the source suspends publication.
• Monthly repricing. Annual volume, monthly price. Shares the risk and keeps the relationship intact.
• Inventory as the hedge. Carrying stock is taking a long position whether or not you call it that. At RMB 71,500/t, a 1,000 t position is RMB 71.5m tied up; at 8% a year that is about RMB 15,700 a day in carry. Our warehousing page works through that arithmetic properly.
• Payment structure. For first orders, an instrument you can actually enforce against. The quotation page covers what each option really costs.