The Gulf sulfur squeeze reaches fertilizers and nickel

On September 3, 2026, Mosaic warned that its Louisiana phosphate operations were likely to remain idled for an extended period, with jobs at Uncle Sam and Faustina due to end later in the year. The company cited sulfur availability and cost. It also made a crucial distinction: ammonia production at Faustina would continue. Mosaic’s Louisiana update.

The difference comes from the production process. Phosphate plants need sulfuric acid to treat the rock. That reagent links fertilizers to oil, gas and parts of the metals industry. When its supply tightens, otherwise unrelated plants face a common constraint.

Ore needs a processing route

Commercial sulfur does not come only from deposits mined for sulfur itself. It is also recovered during oil refining and natural-gas purification. Material removed from hydrocarbons becomes an input for other industries. USGS describes this recovery route separately from mined sulfur. Industrial process background.

Its main end use is the manufacture of sulfuric acid. The two materials are connected, but they are not interchangeable. Sulfur is a feedstock; acid is a processed chemical with particular concentrations, transport requirements and applications. A price per tonne of sulfur cannot simply be compared with a price per tonne of acid. USGS and Sumitomo’s product specification.

In the wet-process route to phosphate fertilizers, acid reacts with phosphate rock to produce phosphoric acid, which is subsequently used to make fertilizer. A deposit of phosphate rock is therefore only part of the production system. The plant also needs the chemical that makes the rock usable. EPA’s technical documentation describes this step and the separation of the resulting gypsum. EPA, section 8.9.2.1.

Some nickel processing faces a similar economic dependency. HPAL, or high-pressure acid leaching, uses sulfuric acid to recover metal from oxide ores. Leaching puts the target metals into solution so that they can be separated. The affected Indonesian plants produce an intermediate known as MHP, mixed hydroxide precipitate containing nickel and cobalt, for the battery-material supply chain. MHP is not a finished battery, and a tonne of it is not a tonne of pure nickel. Process definition and Reuters’ description of the plants.

The dependency map looks different once that intermediate step is visible. A fertilizer producer and a nickel processor can end up bidding for the same sulfur cargoes. They are not competitors in their final-product markets, but they may compete for an essential input.

The squeeze was already there

On January 28, 2026, CRU was already describing a tight sulfur market in 2025, supply interruptions and growing demand from phosphate and nickel production. The analysis also highlighted stock releases that helped offset shortages. Its diagnosis predates the spring conflict; its forecasts must not be mistaken for production subsequently achieved. CRU’s prewar assessment.

There was an industrial warning too. A CRU article published on January 29 reported that Mosaic had announced steps to idle Brazilian single superphosphate production on December 16, 2025, following the rise in sulfur costs. The 2026 disruption reached a value chain that was already under strain. Announcement chronology.

The Gulf crisis added shipping risk to that existing tension. By March 6, Reuters was reporting the exposure of Indonesian processors reliant on Middle Eastern cargoes. This is a dependency on traded flows. It does not mean that all global sulfur production lies behind the Strait of Hormuz, or that every route has been blocked in the same way throughout the conflict. Reuters’ March 6 report.

A plant’s relevant supply is not the amount of sulfur that exists worldwide. It is the material with the right specifications that can reach the site before inventories run out, at a price compatible with continued production. Abundance somewhere else does not keep next week’s production schedule intact.

Mosaic’s higher bill reaches earnings later

As of August 4, Mosaic said Bartow was operating at 40% of its targeted annual operating rate. The denominator matters: this is neither 40% of total US capacity nor a 40% reduction. In Brazil, the company described commodity fertilizer production being idled while animal-feed production at Cajati continued. Second-quarter results.

Reuters had reported the Brazilian measures in July, including extended shutdowns and a plan to begin gradually idling Uberaba in September. The published schedule remains a plan; we have no confirmation of each step completed since. July 8 dispatch, republished July 9.

The financial release reveals a particularly important lag. Average sulfur consumed in the Phosphate segment cost $522 per long ton in the second quarter. Third-quarter contracts with U.S. Gulf Coast refiners were agreed at $705 per long ton, but Mosaic said the new contract would primarily affect fourth-quarter operating results. One figure is a consumed cost; the other is a contract price. They are not two equivalent observations of a spot benchmark. Phosphate Results and Outlook.

This timing matters when prices begin to ease. Purchasing, shipping, inventory consumption and finished-product sales do not happen simultaneously. An existing contract does not automatically reset to the latest market quote. Conversely, material bought more cheaply can temporarily cushion a manufacturer during a price surge.

Lower output creates another problem: some costs remain even as they are spread across fewer tonnes. Mosaic’s accounts explicitly describe idle costs and weaker fixed-cost absorption in the Brazilian segment. They also identify limited customer credit availability in Brazil. The operating squeeze is not reducible to a single, more expensive input. Q2 release, Mosaic Fertilizantes; Form 10-Q, operating review.

Nor can the company’s net loss be relabeled its “sulfur bill.” It also contains foreign-exchange effects, a negative mark-to-market adjustment on its Ma’aden investment, and charges for a project it decided not to proceed with. Those items do not measure a physical sulfur shortage. Form 10-Q, pages 29–30.

A $270-per-tonne sensitivity

A simple calculation shows the scale of the exposure, provided it is not presented as the actual cost borne by every producer.

On July 8, Platts assessed granular sulfur at $1,200 per metric tonne CFR Brazil, compared with $525 on February 26. CFR includes ocean freight to the named destination port, not every subsequent cost of delivering material to a plant. ICC definition. In the same report, Mosaic said that producing ten tonnes of DAP or MAP (diammonium and monoammonium phosphate) required about four tonnes of sulfur. This is the company’s stated input coefficient, not a universal constant for every process. Platts assessments and Mosaic’s statement.

Applying that coefficient to the two prices takes the sulfur contribution from $210 to $480 per tonne of fertilizer, an increase of $270. This is an arithmetic sensitivity with other factors held constant. It is neither total manufacturing cost nor an observed change in the price paid by a farmer.

Calculation: (4 ÷ 10) × 525 = 210; (4 ÷ 10) × 1,200 = 480; 480 − 210 = $270 per metric tonne of DAP or MAP. Inputs published by Platts.

A producer can absorb the increase in its margin, pass some of it to customers, share it elsewhere in the supply chain, or reduce activity. None of those responses guarantees that the same quantity of fertilizer will be produced and used.

The farmer’s next decision also depends on crop prices, soil conditions and available cash. Paying more and adjusting purchases are different responses to the same shock. The mechanism identifies exposure in agricultural margins and volumes; it does not establish a particular percentage loss in crop yields or a predetermined rise in food prices. This investigation does not have the agronomic model or counterfactual needed for such an estimate.

Nickel exposure is process-specific

On April 14, Reuters reported cuts of at least 10% at several Indonesian MHP plants, citing three unnamed sources. Several plants had previously operated above nameplate capacity, and the reductions brought them back toward that level. This was not evidence of a 10% decline in all Indonesian nickel production. It was also a spring observation, not a measurement of September operating rates. Reuters’ reporting.

Other constraints matter. Indonesian mining quotas also affect ore availability and prices. An output cut, by itself, therefore cannot identify how much of the decline was caused by sulfur. Reuters on mining quotas.

An HPAL plant needs a reagent to process its ore. If that reagent is unavailable or prohibitively expensive, an adequate ore supply does not guarantee intermediate production. A higher metal price cannot make a delayed cargo arrive.

Processing routes that do not use this acid-leaching step have a different exposure. Consumption per tonne of recovered nickel also depends on the ore and the plant’s operation. Turning one facility’s input ratio into a coefficient for the whole country would create false precision. HPAL’s technical scope.

In Congo, selling acid supports the smelter

Some operators encounter sulfuric acid from the other direction: they produce it while processing sulfide ores. Sulfur-bearing gases from certain smelters are captured and converted into acid. The chemical becomes a product to sell rather than an input to purchase. EPA, section 8.10.2.

For the second quarter of 2026, Ivanhoe Mines reported 119,603 tonnes of acid sold and $56 million of related revenue at Kamoa-Kakula. These are complex-level figures, not net profit or Ivanhoe’s consolidated revenue. Results released July 29.

The cost table makes the mechanism clearer: smelter operating cost was $0.41 per payable pound of copper produced, against a $0.39 acid credit. Yet total C1 cash cost was still $2.84 per pound, including other operations. C1 is an industry cost measure not defined under IFRS Accounting Standards. C1 cost breakdown.

Acid sales reduce the net cost reported by the company. The credit is already included: adding those receipts to the margin again would count them twice.

This arrangement is not unique to Congo. Acids Co., the Japanese business associated with DOWA and Sumitomo, describes production from copper- and zinc-smelting gases and exports to, among others, HPAL facilities in the Philippines. The arrangement illustrates industrial complementarity. The company does not identify volumes available for new customers. Acids Co.’s supply system.

Why supply takes time to respond

Higher prices can encourage stock releases, remelting, changes in cargo destinations or investment in recovery. CRU described such adjustments before the war. Supply is not completely fixed. Inventories and trade routes.

But when an input is recovered during another activity, a stronger price incentive is not always enough to obtain much more of it quickly. A refinery still has to run its fuel business; a gas plant must process gas; a smelter needs metal concentrates. Finding a more profitable outlet for sulfur or acid does not remove those conditions. This follows from the co-production described by USGS and EPA; it is not a quantified supply forecast. USGS and EPA.

Buying finished acid can bypass one processing step, but it changes the logistics. Transport and storage must be suitable for the chemical. Sumitomo’s product specification distinguishes the concentrations sold and the transport equipment. Sulfur and acid cargoes cannot be substituted simply by editing a purchase order. Product specification.

There is a counterpoint to an uninterrupted price-surge narrative. In a September 1 publication, CRU assessed sulfur CFR Brazil at $1,075 per tonne for the week ending August 27, after an easing from its June and early-July peaks. This is a CRU assessment, separate from the Platts observations used in our calculation; the two are not spliced into a continuous series. A partial decline does not show that every producer’s contracts, inventory costs and margins have already normalized. CRU assessment.

Mineral security starts before the metal

The documents establish production decisions, more expensive inputs and delayed effects on earnings. They do not support a global nickel-output loss attributable to sulfur alone, let alone a certain estimate of resulting food inflation.

The practical lesson is still clear. Securing ore while leaving its processing dependent on a reagent exposed to disrupted shipping is an unfinished supply strategy. Recovery capacity, contracts and inventories are part of the production system, not peripheral purchasing details.

The recovery will have to be traced through deliveries and plants: input availability, returning volumes, the cost of materials actually consumed, and finally margins. Sulfur is a reminder that a metals-supply strategy can fail several stages before the metal appears.

Further reading

This processing constraint complements our investigation into U.S. copper inventories and their carrying costs. For the hydrocarbon side, Banking on Oil follows the contracts, financing and infrastructure behind oil flows.

Sources

Source collection closed on September 12, 2026. Links appear in the text and figure captions. USGS and EPA technical documents support the explanation of processes; their historical statistics are not presented as current market data.

Mosaic decisions and accounts. Louisiana update, September 3; Q2 results, August 4; 10-Q signed August 5, quarter ended June 30. Company statements and unaudited interim accounts, with different reporting scopes.

Prices and the pre-existing squeeze. CRU on January 28 and January 29; Platts, February 26 and July 8 assessments; CRU, week ending August 27, published September 1. CFR is defined by ICC. The two providers’ assessments are not merged.

Processors and operating decisions. Reuters on supply, March 6, mining quotas, March 6, HPAL cuts, April 14, and Brazilian shutdowns, July 8 dispatch. Linked pages are identified republications of Reuters reporting.

Acid and copper. Ivanhoe Mines, July 29 results, quarter ended June 30: volumes sold, revenue and C1 table. Acids Co. describes its own industrial supply network.

Processes. USGS sulfur overview and the compendium’s Sulfur section; EPA 8.9, phosphoric acid and 8.10, sulfuric acid; Sumitomo’s HPAL glossary and acid product sheet. Both EPA sheets date from 1993 and were reformatted in 1995. They explain chemistry, not current capacity.

Limitations

This investigation uses public documents, specialist data and published reporting. l0g conducted no plant visits or interviews. Mosaic’s interim financial statements are unaudited. Corporate explanations remain company statements, including those filed with regulators. Reuters’ April reporting on Indonesian cuts relies partly on unnamed sources; this article does not establish those facilities’ current operating rates. The $270 calculation is a constant-coefficient sensitivity, not an audit of any individual plant. The investigation does not provide an exhaustive September 12 price survey, a measurement of the Gulf’s share of global supply, or a general estimate of crop losses.