South African ferrochrome output is on course to have halved, and then halved again, inside three years. That is a supply contraction of a kind the chrome chain has not had to absorb in living memory, and it is happening in the open, with figures anyone can look up.
The sellers standing closest to that contraction, the chromite miners of Balochistan, cannot see it in any price they are quoted. There is no published assessment of Pakistani chromite. There is no reliable figure for how much of it leaves the country each year. A market is moving in their favour and they are negotiating blind.
The contraction, as far as it can be measured
Fastmarkets, presenting during LME Week in October 2025, put South African ferrochrome production at roughly 3 to 3.3 million tonnes in 2024, and projected to fall to roughly 1 million tonnes in 2026. A figure of roughly 2 million tonnes for 2025 circulates alongside those, and we have not been able to confirm it, so we do not rely on it. South Africa accounts for more than half of global ferrochrome output, so those are not national figures with a local consequence. They are the world number.
Forecasts deserve scepticism, and the 2026 range is a forecast. The 2024 and 2025 figures are not, and they are corroborated by what individual producers have reported. Glencore's joint venture attributable output fell 28 per cent, from 599 kilotonnes in the first half of 2024 to 433 kilotonnes in the first half of 2025. The Glencore and Merafe chrome venture announced temporary suspensions at three smelters, Boshoek with effect from 1 May 2025 and Lion and Wonderkop from 31 May 2025. Lion has since resumed partial production after maintenance, so this should not be read as three plants permanently down.
Idled smelters are the detail worth sitting with. A ferrochrome furnace is not a tap. Restarting one is a capital event with a lead time measured in quarters, and the decision to restart is taken against forward power costs rather than today's alloy price. Capacity that goes cold in a downturn does not reappear the moment the market turns. That is the mechanism by which a demand recovery in stainless steel, whenever it comes, meets a supply base that cannot answer it quickly.
The benchmark went the other way
Here is where an honest account has to slow down.
The European benchmark for the third quarter of 2026 was set at 156 US cents per pound of chrome by Samchrome FZE, as reported by SMM and Metal.com. The second quarter 2026 benchmark was 161 cents. That is a fall of roughly 3 per cent on the quarter. We note that some trade coverage characterised the same settlement as a rise, which usually indicates a different basis or currency, and is worth resolving before either direction is quoted at a buyer.
A falling benchmark alongside collapsing output is not a contradiction, but it is a warning against reading the production data as a straightforward bull case. The European benchmark is a negotiated quarterly settlement between a producer and its consumers. It prices the alloy, not the ore, and it reflects the balance of the alloy market at the moment the negotiation closes, including inventory, contract volumes and the state of stainless demand. A settlement is a bargain struck between two parties, not an observation of a cleared market.
We would not present the Q3 figure as evidence that the supply data is wrong, and we would not tell a seller that the supply data guarantees a higher price next quarter. What the two together establish is that the alloy market has so far absorbed a very large loss of South African capacity without repricing upward.
The feed price, and its shelf life
The number closer to a chromite seller's own position is the concentrate assessment. South African 40 to 42 per cent chrome concentrate was assessed at 295 to 305 dollars per tonne in bulk on 10 March 2026, according to Mining Bulletin.
That is the most recent assessment for a comparable feed material we are prepared to publish, and it is five months old at the time of writing. We flag the date deliberately. A concentrate assessment quoted without its date, in a market where the alloy benchmark has moved between quarters, is the kind of figure that ends up inside a landed cost model and stays there long after it has stopped being true.
What Pakistan actually has in the ground
The geology is well described. Peer reviewed work on the Muslim Bagh ophiolite reports chromium grades running from roughly 30 to 60 per cent and averaging near 44 per cent, with separate beneficiation work upgrading 38 per cent Cr2O3 feed to about 60 per cent. Muslim Bagh and Khanozai together account for more than 95 per cent of national output, a share that is widely reported though we have not traced it to a named issuing body, with further occurrences at Kila Saifullah, Zhob, Lasbela and Khuzdar, and minor deposits in Khyber Pakhtunkhwa.
A 28 per cent product and a 56 per cent product are not the same commodity, and the difference between them is the whole of a seller's margin. A range that wide inside two adjacent districts means grade is a property of the individual lot rather than of the origin. That is worth stating because buyers routinely price Pakistani chromite as an origin rather than as an assay, and an origin price is always set at the bottom of the range the origin is known to produce.
The number nobody can give you
There is no independent assessed price for Pakistani chromite. We looked. The only 2026 figures we could find were trader listings, which are offers rather than assessments, published by parties with a position in the outcome.
This is not a technicality. An assessment is a third party's account of where material actually cleared, published on a date, by an organisation whose business is the credibility of that account. A listing is somebody's asking price. When no assessment exists for a material, price discovery does not vanish. It relocates. It happens entirely inside the buyer's information advantage, because the buyer sees a hundred lots a year from a dozen origins and the seller sees his own.
The absence compounds a second gap. Production estimates for Pakistani chromite range from 20,000 to 237,000 tonnes a year across the sources we examined. That is a spread of more than ten to one on the most basic question anyone would ask about a market. We are not going to pick a number out of that range and present it as a figure. Neither should anyone else, and a report that does so without showing the spread is telling you something about its methodology rather than about Pakistan.
It appears that essentially all Pakistani chromite is exported raw to China, and that there is no viable domestic ferrochrome capacity. We hold that with medium confidence and would want it corroborated before it went into a contract assumption, but it is consistent with everything else we can see.
Negotiating without a reference
A seller with no assessment for his own material is not without options, but the options are contractual rather than commercial.
The first is to price against something visible. A contract that references a published assessment for South African concentrate of a stated grade, with an agreed and written discount for grade differential and freight, is a contract whose terms can be checked next quarter. A contract that references a negotiated figure with no published anchor cannot be checked at all, which is precisely why it is offered.
The second is assay discipline. Where the origin's grade range runs from roughly 30 to 60 per cent, the only defence against being paid at the bottom of it is a certificate from a laboratory the buyer has agreed to in advance, taken at a sampling point both parties witnessed. This costs very little and is the single largest available uplift on tonnage a seller is already shipping.
The third is to understand what the buyer is hedging. A Chinese smelter buying Balochistan chromite in 2026 is buying into an alloy market that has lost a great deal of South African supply and has not yet repriced. He is more exposed to that than his offer suggests. The seller who knows the Fastmarkets output figures and the quarterly benchmark history is negotiating with roughly the same picture of the world that the person across the table has, which is a different conversation from the one that happens when only one side has been reading.
The supply shock is real and it is documented. What is missing on the Pakistani side of it is not geology, and not demand. It is a price anyone can point at.