Global Phosphate Rock & Yellow Phosphorus Market Size, Share & Forecast 2026–2033

Scope: the four per cent of the rock that makes every non-fertiliser phosphorus molecule

World phosphate rock mine production was 250 million tonnes in 2025, up from 239 million the year before, by the U.S. Geological Survey’s February 2026 estimate. Almost all of it becomes fertiliser. The share that does not — the tonnage that goes instead into an electric furnace and comes out as elemental phosphorus — is about 9.2 million tonnes, or 3.70 per cent of the total. Every non-fertiliser phosphorus molecule in commercial existence starts there: the phosphoric acid in a soft drink, the flame retardant in a cable, the scale inhibitor in a cooling tower, the glyphosate on a verge, the PCl₃ that becomes a lithium battery electrolyte salt.

That asymmetry is the whole reason this report exists as something other than an agri-commodity note. A market of 250 million tonnes priced in dollars per tonne of rock is a fertiliser market. A market in which four per cent of the tonnage carries 8.70 per cent of the value and one hundred per cent of the industrial chemistry is a different question, and it is the question a buyer of phosphorus compounds actually has to answer.

Two pools are counted. The first is phosphate rock at market value — world mine production valued at the realised price, not the fertiliser contract price, so that the same tonne is not counted twice downstream. The second is elemental phosphorus, counted at its value added: the yellow phosphorus price less the rock cost embedded in it. The arithmetic of that netting is set out in the value chain chapter and in the risk chapter, because it is the single largest modelling choice in this report and a reader is entitled to see it.

Four things are excluded, and each is a decision. Purified wet-process phosphoric acid sold to cathode makers is a separate market with a separate report in this segment: it is made from rock, but it is not made through the thermal branch, and its economics are set by purification yield rather than by a power tariff. Merchant phosphonates and merchant phosphorous acid are likewise a separate report, because a phosphonate tonne and a phosphorus tonne have different buyers. Downstream phosphate salts — STPP, TCP, feed-grade MCP and DCP — are counted only to the extent that their rock and thermal inputs sit inside the two pools above; their own conversion margins are not. And phosphate fertiliser prices are used throughout as a comparator, never as a component of the forecast, because a DAP price moves on a crop cycle and this market does not.

The perimeter that survives is the physical base of the entire phosphorus complex: the rock, and the one branch of it that a mine cannot serve directly.

The number

Read the last row before the first. Value compounds at 3.98 per cent against 2.52 per cent for rock tonnage, so realised revenue per tonne of rock rises from USD 129.24 to USD 144.79 — +12.03 per cent in eight years, roughly 1.4 per cent a year. That is a modest premium, and its modesty is deliberate: phosphate rock is not a scarce commodity. World reserves are 73 billion tonnes and identified resources exceed 300 billion. There is no imminent shortage of the stuff, and a forecast that assumed one would be a forecast about geology that the geology does not support.

The premium comes from mix rather than from scarcity. Elemental phosphorus grows at 2.92 per cent against 2.52 per cent for rock, and its value added per tonne is roughly twenty times rock’s own price, so a small shift in the tonnage split moves aggregate value more than a large move in the rock price does. Everything else in this report is an examination of whether that 2.92 per cent is achievable — and the answer turns on a capacity number that is being reduced by policy at the same time as demand for it is rising.

The five pools

The table is an allocation of the USD 32,310.0 million to its end use, not five independent market sizes, because the rock underneath several of them is the same rock. What it shows is a market that is four-fifths fertiliser by value and eighteen per cent everything else — and that the eighteen per cent is where the pricing power, the technical barrier and the entire investment case sit.

Note the third row against the second. Battery-grade purified acid is already 4.09 per cent of the value of a market whose thermal branch is 8.70 per cent, and it is the only pool in the table whose volume driver is a forecast about cars rather than a forecast about people eating. That is why the cathode chain is treated separately in this segment and why the reader should not add the rows together to size it.

Regional structure

Two features of this table matter more than the ranking. The first is that the country with the largest reserves is not the country with the largest production: Morocco holds 68 per cent of the world’s declared reserves and produces 14.4 per cent of its rock, while China produces 44 per cent on 4.7 per cent of the reserves. Reserve life at current production is roughly 1,390 years in Morocco against 31 in China and 50 in the United States. The second is that the United States is a net importer with net import reliance of 16 per cent, up from 11 per cent four years earlier, because Florida reserves are depleting and the grade is falling. Phosphate rock was added to the U.S. Final 2025 List of Critical Minerals on 7 November 2025.

Forecast snapshot

The forecast assumes no new greenfield elemental phosphorus capacity of scale outside China, Chinese effective capacity stabilising in the 1.18 to 1.22 million tonne band, and rock capacity additions in Brazil, Kazakhstan, Mexico, Morocco and Russia completing on the schedule the USGS records for 2028. It assumes P₂O₅ fertiliser consumption rising from 47.8 to 51.5 million tonnes by 2029, and it assumes the current rock price recovery holds rather than extends.

Report structure

1. Value chain — eight tonnes of rock, thirteen megawatt-hours, and one molecule a mine cannot ship
2. Technology and innovation — the submerged-arc furnace is the technology, and it has not changed in fifty years
3. Competitive landscape — an administered denominator, not a competitive one
4. Regulatory landscape — capacity is being closed by policy, not by economics
5. Investment activity — the money is buying the customer, not the furnace
6. Market risk — four ways this forecast is wrong, and the one I would watch

Each chapter is written to be read on its own, and each carries its own falsifiers rather than deferring them to a conclusion.

Companion analysis. The downstream of this market is written up in this segment in One Oxygen Is the Whole Market Boundary, which sets the P(III) branch against the P(V) branch and shows why 58 per cent of a tonne of HEDP is the yellow phosphorus inside it — the mechanism by which a furnace price becomes a specialty chemical price.

Table of Contents

1. Value chain
2. Technology and innovation
3. Competitive landscape
4. Regulatory landscape
5. Investment activity
6. Market risk

Competitive Landscape

Competitive landscape: an administered denominator, not a competitive one

The conventional way to describe competition in yellow phosphorus is to count capacity, rank producers and talk about utilisation. That framework fails here, because the number that everything divides by is being set by policy rather than by investment, and a utilisation rate computed against a shrinking denominator is not a competitive statistic.

The facts are these. Chinese installed yellow phosphorus capacity is approximately 1.58 million tonnes across four provinces — Yunnan, Sichuan, Guizhou and Hubei. Actual 2025 output was about 1.0 million tonnes, a 63 per cent utilisation rate on nameplate. Global capacity is approximately 1.8 million tonnes, of which China is about 85 per cent, with the balance in Morocco, the United States and Kazakhstan — mostly aging lines, with a global operating rate around 55 per cent in May 2026 and insufficient overseas tonnage to fill a Chinese shortfall.

Now the denominator. Under the dual-control energy regime, tighter environmental enforcement and stricter safety protocols, new Chinese capacity is being held back and obsolete capacity is being retired, with effective capacity expected to stabilise in a 1.18 to 1.22 million tonne band through 2026. That is a 24.05 per cent reduction against installed nameplate. And here the arithmetic turns: at 1.0 million tonnes of output against 1.20 million tonnes of effective capacity, the operating rate is 83.33 per cent — not 63 per cent. Both numbers are true. They describe different questions, and only one of them is the question a buyer should ask.

What the monthly data actually shows

The utilisation series is a series about weather, tariffs and accidents, and it moves violently. On 13 March 2026, Longzhong recorded weekly output of 17,094 tonnes at 54.49 per cent utilisation, with Yunnan at 50.7 per cent, Guizhou at 53.9 per cent and Sichuan at 52 per cent. By 8 October 2026, weekly output was 22,140 tonnes at 67.36 per cent, with Guizhou at just 49.7 per cent after a mainstream producer cut back. A single plant event moves the national rate by points.

That is not a competitive market in the textbook sense. It is a market where supply is a residual of something else — hydroelectric availability in the southwest, a provincial energy-consumption quota, a maintenance decision at a plant that is the only local employer — and where price is the clearing mechanism for the residual. The firms that do well are those whose cost position is structurally below the marginal furnace, which in practice means those with captive rock, captive hydro or an off-gas recovery that converts electricity into a second revenue line.

The structure at the top

Ownership in China is concentrated in a small number of provincial champions and integrated fertiliser groups, typically state-linked, for whom phosphorus is one link in a chain that runs from mine to phosphate fertiliser to fine chemicals. That integration is the competitive moat, because it lets a group run a furnace through a low-price year without shutting it, and it explains why merchant buyers of P₄ have less leverage than the concentration numbers would suggest.

Outside China the picture is thinner than the capacity figures imply. Morocco’s elemental phosphorus position is small relative to its rock position, and the country’s strategy is explicitly to convert rock into higher-value phosphates rather than into phosphorus metal. The United States still converts part of its rock into elemental phosphorus for industrial compounds, principally the glyphosate chain, but the volume is a fraction of the historical base. Kazakhstan’s capacity is ageing. The consequence is a market with a single dominant supplier and no credible swing producer, which is the structural reason the 2026 price response to a 40,000 tonne outage was a 23.78 per cent move.

What competition turns on

Not brand, not specification and not service — those matter downstream, not here. It turns on three things: the delivered cost of electricity, the security of a rock supply at a grade the furnace can take, and the regulatory permission to run. A producer can be the lowest-cost operator in the country and still lose volume to a plant with a better tariff and a friendlier provincial quota. That is why capacity in this market is best read as a political variable, and why the forecast in this report treats Chinese effective capacity as a policy output rather than a market one.

Value Chain

Value chain: eight tonnes of rock, thirteen megawatt-hours, and one molecule a mine cannot ship

The chain begins with a number most market reports never write down: how much rock it actually takes to make a tonne of elemental phosphorus. The stoichiometry is fixed and not open to opinion. Phosphorus pentoxide carries 61.95 grams of phosphorus per 141.94 grams of P₂O₅, so one tonne of contained phosphorus is 2.2912 tonnes of P₂O₅. At the 30 per cent P₂O₅ concentrate grade that a furnace is actually fed, that is 7.637 tonnes of rock; allowing an eight per cent loss across beneficiation, furnace dust and slag, the working figure is 8.248 tonnes of rock per tonne of P₄. Every cost claim in this chapter sits on that number.

The second input is electricity, because the reduction is carbothermal and the bond being broken is not cheap. A modern submerged-arc furnace consumes roughly 13,000 kilowatt-hours per tonne of phosphorus, alongside about 1.6 tonnes of coke or anthracite as reductant and a silica flux. The furnace runs continuously, at a power draw of tens of megawatts per unit, and it is therefore not a chemical plant in the ordinary sense: it is a power contract with a reactor attached.

The cost stack, in one tonne

Take the March 2026 Chinese market, where the domestic yellow phosphorus price was RMB 26,200 per tonne, or USD 3,690.14 at RMB 7.1. Rock at USD 143 per tonne — the delivered industrial basis this segment uses consistently across the phosphate reports — puts the rock bill at USD 1,179.46, or 31.96 per cent of the selling price. Power at RMB 0.42 per kilowatt-hour puts the electricity bill at RMB 5,460, or USD 769.01, 20.84 per cent of the price. Together, rock and power are 52.80 per cent of the invoice before a single electrode, refractory, wage, effluent consent or financing cost is paid.

That leaves 47.2 per cent for everything else, and the market tells us how much of it survives. Longzhong’s assessment for 13 March 2026 puts industry theoretical profit at RMB 4,305 per tonne on a RMB 26,200 price — a 16.37 per cent margin. So the non-rock, non-power cost base is about 30.8 per cent of price: reductant, electrodes, maintenance, labour, environmental compliance and depreciation. A producer without its own furnace and its own power contract is buying the two largest inputs at spot and living inside that 16 per cent.

Why the mine cannot capture it

A phosphate mine can sell rock to a fertiliser plant, and it can sell rock to a furnace. It cannot ship elemental phosphorus without building the furnace, because phosphorus is not a product that exists in rock in any transportable form — it has to be reduced. That is the structural break in this chain and the reason the thermal branch has pricing behaviour that rock does not: the rock price is set by a buyer base of hundreds of fertiliser plants, and the phosphorus price is set by a small number of furnaces whose marginal cost is an electricity tariff.

Which produces the odd result visible in 2026. Through 2024 and 2025 the World Bank’s Morocco rock benchmark (70% BPL, FAS Casablanca) sat flat at USD 152.50 per tonne. By August 2026 it was USD 170.00, up 11.48 per cent year on year. Yellow phosphorus over the same stretch moved far more — from RMB 26,200 in mid-March 2026 to a benchmark of RMB 32,429.33 on 19 May, a 23.78 per cent move in nine weeks and 44.16 per cent year on year. The rock is not what moved. The power was, and a fire at a Guizhou plant that took 40,000 tonnes and four furnaces offline on 14 May turned a tightening market into a squeeze.

Where the chain actually leaks value

Downstream, the thermal branch splits three ways. Thermal phosphoric acid — made by burning phosphorus and hydrating the oxide — was quoted at RMB 8,000 per tonne for 85 per cent industrial grade in Jiangsu in March 2026. Phosphorus trichloride was RMB 6,700, and phosphorus oxychloride RMB 6,350. Run those against the P₄ price and the conversion margins are thin in the acid step and thinner in the chlorination step, which is why the value in this chain does not sit in the first conversion. It sits in the third and fourth, where a molecule acquires a specification, and that is a different report in this segment.

The chain therefore has a shape worth stating plainly: a mine that cannot move without a furnace, a furnace that cannot move without a tariff, and a chemical industry that owns neither but captures most of the margin. The forecast in this report is a forecast about the middle link.

Investment Activity

Investment activity: the money is buying the customer, not the furnace

Almost nothing is being built in this market, and that is the investment case rather than a gap in it. A greenfield submerged-arc phosphorus furnace is a large, single-purpose, electricity-hungry asset in a jurisdiction that is actively restricting energy-intensive industry, and no Western capital is proposing one. The capital that is moving goes into two other places: rock capacity in jurisdictions that want it, and downstream conversion businesses that own a specification.

Where the capital is actually going

On the rock side, the capacity additions the USGS records for completion by 2028 are in Brazil, Kazakhstan, Mexico, Morocco and Russia, with projects for the following decade under development in Canada, Congo (Brazzaville), Guinea-Bissau and Senegal. These are mine and beneficiation investments, and their logic is fertiliser, not phosphorus metal. On the capability side, the notable movement is into purification: the ability to take wet-process acid to battery grade, which is where this segment’s purified acid report sits, and which is where a rock owner can capture a cathode-chain premium without building a furnace.

On the industrial side, the pattern this segment has documented downstream repeats here. Buyers are paying for validated customer positions rather than for tonnage, because tonnage is available from a hundred plants at a published quote. In the phosphonate market that produced a 1.86 times sales, 10.3 times EBITDA acquisition and a reported USD 1.6 billion negotiation for a specialty platform. In this upstream market the equivalent move is a fertiliser or chemicals group acquiring an integrated phosphorus producer for its rock, its power contract and its permits together — the three things that cannot be bought separately at a predictable price.

Why the capital intensity repels entrants

The capital cost of a tonne of phosphorus nameplate is high, the permitting timeline is long, and the asset is only economic against a power price that the operator does not control. In the wet-process acid business the estate has already priced this problem: a 280,000 tonne purified-acid project at RMB 4,264 per tonne of nameplate against a gross spread of RMB 2,902 is 1.47 years of its own spread, which is a defensible project only while the spread holds. A phosphorus furnace is a harsher version of the same arithmetic, because its cost is concentrated in a single input with a published tariff.

The result is a market where the marginal supplier is an existing asset, where brownfield debottlenecking and off-gas recovery are the realistic capacity additions, and where a demand forecast that requires new capacity is effectively a forecast that requires a policy reversal.

What an investor should watch

Three signals, in order. First, the Chinese effective-capacity figure: if it stabilises above 1.22 million tonnes rather than inside the band, the supply thesis in this report weakens and the revenue-per-tonne premium goes with it. Second, the spread between the yellow phosphorus price and the thermal acid price — a widening spread means the furnace is capturing the constraint, a narrowing spread means the downstream is absorbing it, and only the first supports the value path here. Third, capital announcements in purified wet-process acid: every tonne of battery-grade acid capacity that is commissioned reduces the pull on the thermal branch from the cathode chain, and the thermal branch is the entire premium in this forecast.

Technology & Innovation

Technology: the furnace is the technology, and it has not changed in fifty years

There is a temptation, in a market this old, to treat the technology chapter as a formality. It is not, because the absence of innovation here is the reason the supply curve is shaped the way it is. Elemental phosphorus is made by charging phosphate rock, silica and carbon into a three-electrode submerged-arc electric furnace, driving the reduction at around 1,400 to 1,500 degrees Celsius, and tapping phosphorus vapour which is condensed under water. The slag leaves as calcium silicate and is sold or dumped. The off-gas is mostly carbon monoxide and is either flared or, in the better plants, recovered.

Nothing about that description would be out of place in a 1970s process manual, and that is not because nobody has tried. It is because the chemistry offers no alternative. The carbothermal reduction of phosphate is thermodynamically brutal and electrically intense, and every proposed route to elemental phosphorus that avoids the furnace has foundered on the same three rocks: energy, scale and the fact that phosphorus has no ore mineral that can be separated mechanically.

What has actually improved

Three things, and each is an efficiency rather than a route change. The first is furnace scale and electrode practice: larger units and better electrode management have pushed specific energy consumption from the 14 to 15 megawatt-hour band toward 13, which is a seven to fourteen per cent saving on the single largest cost line — worth more than most process innovations in this industry. The second is off-gas recovery, which converts a flared carbon monoxide stream into either thermal energy or a chemical feedstock, and which is the main route by which a Chinese producer can move its effective power cost. The third is feedstock flexibility on grade, where the practical constraint is not P₂O₅ content but the ratio of calcium oxide to silica, because the slag has to flow. That is why a furnace can sometimes run lower-grade rock more cheaply than a wet-process acid plant can, and why the two branches compete for different parts of the same deposit.

The wet route does not compete, and that is the important point

It is worth being precise about why, because the mistake is common. Wet-process phosphoric acid produces P(V) — phosphate — and there is no commercial path from phosphate back to elemental phosphorus that does not go through a furnace. A producer who wants a phosphite, a phosphonate, a PCl₃-derived electrolyte salt, an organophosphorus flame retardant or a thermal-route purified acid cannot begin from wet acid. The two branches are not substitutes competing on cost; they are two different products that happen to share a mine.

The consequence for technology is that a cost breakthrough in the wet route does nothing for this market, and a cost breakthrough in the furnace would reprice a large slice of industrial chemistry at once. That asymmetry explains why this is one of the few chemical sectors where the incumbent route has attracted so little disruptive investment: there is no alternative to disrupt it with, so capital goes instead into making the incumbent cheaper to run and into securing its inputs.

The three real technical constraints, in order of severity

Electrode supply. Graphite electrodes are shared with the electric-arc steel industry, which is a far larger buyer, and electrode prices transmit steel-cycle pressure straight into a phosphorus producer’s cost base. Effluent and slag. Elemental phosphorus production generates phosphorus-bearing wastewater and a calcium silicate slag; in China, discharge consents and slag handling are now the practical limit on utilisation at several plants, and the dual-control energy-consumption regime adds a second constraint that is administrative rather than physical. Safety. White phosphorus ignites spontaneously in air and is acutely toxic; storage, transport and the condensation step carry a risk profile that makes insurers and local regulators cautious, which raises the effective cost of capacity in exactly the jurisdictions where demand is growing.

The honest summary is that this market’s technology story is a story about the cost of electricity and the cost of compliance, and that both are set outside the industry. A reader looking for a technology-driven forecast — a step change in yield or a novel route — will not find one here, and the forecast in this report does not assume one.

Market Risk

Market risk: four ways this forecast is wrong, and the one I would watch

The forecast is USD 32,310.0 million in 2025 growing 3.98 per cent to USD 44,160.0 million in 2033, on rock volume growth of 2.52 per cent and elemental phosphorus growth of 2.92 per cent. Four things would break it.

Risk one: the power price is the forecast, and I do not forecast it

Electricity is 20.84 per cent of the yellow phosphorus price and the marginal cost of the marginal furnace, so the value path in this report is a smoothed view of a variable this report does not model. The evidence for how violent that variable is sits in the 2026 series: RMB 26,200 per tonne on 13 March, RMB 32,429.33 on 19 May — +23.78 per cent in nine weeks — and back to RMB 27,600 to 28,900 by 8 October. A forecast that assumed a stable price would be a forecast about hydroelectric rainfall in Yunnan. The 3.98 per cent value CAGR should be read as an eight-year average around a series that will spend most of its life away from the average.

Risk two: the capacity denominator is a policy variable, and policy can reverse

The 24.05 per cent reduction from 1.58 million tonnes installed to a 1.18 to 1.22 million tonne effective band is the load-bearing assumption. If a provincial government decides that a large furnace is too valuable an employer and taxpayer to retire, the band widens, utilisation falls against nameplate, and the price premium compresses. This is not a low-probability scenario — it is the normal behaviour of an administratively managed industry over a five-year horizon. The falsifier is concrete: watch the announced effective-capacity figure each year, and if it stabilises above 1.22 million tonnes, mark the premium down.

Risk three: the value-added construction understates the double-count problem

Elemental phosphorus is counted at its price less the rock cost inside it — USD 3,690.14 less USD 1,179.46, or USD 2,510.68 per tonne, times 1.12 million tonnes, giving USD 2,812 million. That avoids counting the same rock twice. It does not avoid every overlap: where a producer is integrated from mine to furnace, the rock’s margin is inside the phosphorus price and therefore inside the value-added figure, and the USD 29,500 million rock pool is a market value imputed at USD 118 per tonne rather than an observed transaction. The honest band around the total is roughly plus or minus 8 per cent, and the CAGR is more robust than the level: shifting the base by 8 per cent moves the CAGR by about 0.1 of a point.

Risk four: the demand side is the fertiliser side, whatever this report says about chemistry

81.71 per cent of the value in this market is fertiliser and feed phosphate. The industrial branch — the reason to read this report rather than a crop report — is 8.70 per cent of value and 3.70 per cent of tonnage. So a collapse in fertiliser affordability is not a side event: world P₂O₅ consumption in fertilisers is projected to rise from 47.8 million tonnes in 2025 to 51.5 million tonnes by 2029, and if that growth fails, rock prices fall, rock production is curtailed, and the aggregate value line in this report breaks even while the industrial branch does everything the forecast says it will.

That is the honest structural risk and it is worth stating without hedging: this report’s headline growth rate is a fertiliser number with an industrial premium attached. An investor with exposure to the industrial chemistry should read the 2.92 per cent phosphorus volume line, not the 3.98 per cent value line, because the first is the part that is genuinely scarce and the second contains 82 per cent of something that is not.

The indicator I would watch is not the rock price. It is the thermal phosphoric acid operating rate in Jiangsu, because it is the clearest single read on whether the furnace constraint is biting. While it sits near capacity, the premium in this forecast is real; when it drops, the market has found its escape valve and the premium is the first casualty.

Regulatory Landscape

Regulatory landscape: capacity is being closed by policy, not by economics

The most consequential regulation in this market does not regulate the product. It regulates the electricity, the effluent and the permission to operate, and its effect is to remove capacity from the denominator while demand for the output keeps rising. That combination is the entire supply thesis.

China: dual control, environmental enforcement and safety

Three regimes act together. Dual control of energy consumption caps both the total energy a province may use and the energy intensity per unit of output, which makes an energy-intensive furnace a liability to the province that hosts it and gives local government a direct interest in restricting its running hours. Environmental enforcement targets phosphorus-bearing wastewater, slag handling and dust, and has become the practical ceiling on utilisation at several plants. Safety protocols around white phosphorus handling have tightened after incidents, adding inspection and shutdown risk.

The stated intent is not to eliminate the industry but to shrink the nameplate to something the provinces are willing to power: effective capacity in a 1.18 to 1.22 million tonne band, against 1.58 million tonnes installed. The 2026 supply plan is therefore a plan for a 24.05 per cent reduction in the denominator, achieved through retirement and non-renewal rather than through a single decree, which means it is difficult to date precisely and easy to under-forecast.

The United States: criticality without a supply response

On 7 November 2025, the U.S. Final 2025 List of Critical Minerals added phosphate rock, alongside boron, metallurgical coal and uranium. The designation is a statement about import reliance — 16 per cent of apparent consumption, up from 11 per cent four years earlier — and about the strategic discomfort of a fertiliser and industrial base whose domestic rock is depleting in Florida and declining in grade. What it is not, yet, is a supply response: the permitting and capital cycle for a new phosphate mine is measured in a decade, and in October 2025 the Bureau of Land Management approved a new mine in Caribou County, Idaho explicitly to replace one that will be depleted within the next decade rather than to add to the base.

Morocco and the reserve question

Morocco holds roughly 68 per cent of declared world reserves and produces 14.4 per cent of world rock. Its policy has been to move down the value chain — into phosphoric acid, into fertiliser, into purified and battery-grade products — rather than to expand raw rock exports. For this report that matters because a country that converts its own rock internally captures the thermal-branch premium only if it builds furnaces, and the announced strategy points the other way. The reserve concentration is therefore a geological fact that does not automatically become a phosphorus-supply fact.

Trade policy and the export question

Phosphorus moves between countries as rock, as acid, as finished fertiliser and, in smaller volumes, as elemental phosphorus and its chlorides. Each of those flows sits under a different tariff and quota regime, and the regimes change with food-security politics rather than with chemical-market logic. The practical implication for a buyer is that security of supply in this market is not purchasable on the spot market at a predictable premium: it requires either a contract with a producer whose cost position survives a tariff change, or ownership of the conversion step. This report does not forecast tariff outcomes; it treats the current regime as the base case and flags a change as a risk rather than a scenario.

What regulation does to the forecast

Two channels. The first is the supply channel just described, which is the reason elemental phosphorus volume is forecast to grow 2.92 per cent a year against rock’s 2.52 — supply is being administratively constrained while demand is not. The second is the substitution channel, and it cuts the other way: total-phosphorus discharge limits in the European Union and increasingly in Asia do not distinguish between a nutrient and a scale inhibitor, so a tightening consent regime is a genuine threat to the downstream phosphonate and detergent pools. The forecast assumes the supply channel dominates to 2033. If the substitution channel accelerates in the largest water markets, the volume line is the first casualty and the value line follows with a lag.