South African Manganese:
More Produced, Less Sold
Apparent production growth,
but marginal supply is being cleared out.
| Key Takeaways ● South32’s September quarter manganese production in South Africa increased by 11% quarter-on-quarter (qoq) to 588,000 wet metric tons (wmt), while sales decreased by 22% qoq to 460,000 wmt (both figures reflect South32’s 54.6% equity interest). ● The decline in sales has two causes: the company actively reduced high-cost road transport volumes, and one vessel shipment was deferred to the December quarter; thus, not all of the sales drop can be attributed to the cut in road transport. ● South Africa’s manganese ore exports are estimated at approximately 26 million tons in 2025, with Transnet rail handling only about 16 million tons, leaving roughly 10 million tons to be trucked to ports. ● Based on parameters disclosed by Exxaro, the landed cost per ton via road transport is approximately 14% higher than that via rail. Back-calculating from November quotes, the road-transported ton is nearing its break-even point (estimated). ● Mineral Cube’s assessment: Road-transported tons are becoming the marginal cost line for South African ores. There is limited room below the 33 RMB/ton-degree level for semi-carbonate at Tianjin Port, but a port inventory of 6 million tons means that clearing out will take time. |
|---|
01 Trigger
Production up 11%, Sales down 22%
On October 19, South32 released its operational report for the September 2026 quarter (Q1 FY2027). Data for South African Manganese (in which South32 holds a 54.6% stake) showed an unusual divergence: under the equity interest basis, quarterly sellable production reached 588,000 wmt, a 11% increase qoq and a 7% year-on-year (yoy) increase; however, sales were only 460,000 wmt, a 22% decrease qoq and a 16% yoy decrease. The production growth stemmed from improved efficiency at the underground Wessels mine, representing normal improvement on the mining side.
Why did sales drop so significantly? South32 provided two reasons in its report: first, “reduced road transport volumes in response to market conditions,” and second, “one vessel shipment was deferred to the December quarter.” Attributing the entire sales decline to the cut in road transport requires slight correction: the deferral of shipping schedules is a timing factor that will be replenished in the next quarter. What truly warrants attention is the first point. A mine operator actively reducing shipments while increasing production indicates that it judges the additional output to be unprofitable at current prices.
| Wet Metric Tons (wmt) | Q1 FY2026 | Q4 FY2026 | Q1 FY2027 |
|---|---|---|---|
| South African Manganese Production | 55.1 | 52.8 | 58.8 |
| South African Manganese Sales | 54.8 | 58.6 | 46.0 |
| Production-Sales Gap | 0.3 | -5.8 | 12.8 |
| Sales Grade (Mn%) | 38.3 | 38.5 | 38.6 |
| Australian Manganese Production | 85.4 | 78.2 | 90.2 |
| Australian Manganese Sales | 94.4 | 92.1 | 86.1 |
Note: South African Manganese figures are based on a 54.6% equity interest; Australian Manganese figures are based on a 60% equity interest; the production-sales gap is calculated by Mineral Cube.
Converted to a 100% basis, South African Manganese produced approximately 1.077 million wmt this quarter and sold 842,000 wmt, resulting in an accumulation of approximately 230,000 wmt of stockpiled goods at mines and inland yards over the quarter (estimated). South32 maintained its FY2027 guidance for South African Manganese production at 2 million wmt unchanged, but the wording is now “continuing to monitor market conditions and optimize the use of high-cost road transport.” This phrasing has been appended to the guidance since August 2025—the premise for the production guidance was “continued use of high-cost road transport”—but the company is now treating this premise as an adjustable valve.
For comparison, in the same quarter, Australian Manganese (GEMCO) production increased by 15% qoq, while sales decreased by 7%, primarily influenced by post-rainy season shipping rhythms. While sales declined in both regions, the causes differed: Australia’s decline was seasonal, whereas South Africa’s was cost-driven.
02 Background
The Transportation Ledger Behind 26 Million Tons
South Africa is the world’s largest producer and exporter of manganese ore. Citing data from the South African Revenue Service (SARS), the Minerals Council South Africa reported that South Africa’s manganese ore exports reached approximately 26 million tons in 2025, setting a new record, compared to 22.3 million tons in 2024; meanwhile, manganese ore production was approximately 20.8 million tons, up from 19.8 million tons in 2024. Approximately 95% of exports went to Asia, with China accounting for about 67% (Minerals Council scope), while Reuters cited a figure close to 68%; India accounted for approximately 16%.
| Indicator | 2024 | 2025 |
|---|---|---|
| Manganese Ore Production (10k tons) | 1980 | 2080 |
| Manganese Ore Exports (10k tons) | 2230 | ~2600 |
| Export Share to China | — | ~67% |
| Export Share to India | — | ~16% |
| Rail Volume (10k tons) | — | ~1600 |
| Road Volume (10k tons) | — | ~1000 |
Note: Export volume exceeding production reflects inventory drawdowns and differences in statistical bases (wet/dry tons, statistical timing); rail and road volumes are approximate; another statistic for China’s share is close to 68%.
The issue lies in transportation. The Kalahari Manganese Field in the Northern Cape Province is approximately 1,000 kilometers from the coast, where rail transport would normally be used. However, Transnet rail currently handles approximately 16 million tons of manganese ore exports annually, with a target of 16.57 million tons for the 2025/26 financial year. This means that of the 26 million tons in exports, approximately 10 million tons—nearly 40%—must be trucked to ports (estimated). The record export levels in 2025 were driven not by rail capacity expansion, but by trucks filling the gap.
Filling the gap with trucks comes at a cost. In June this year, Exxaro disclosed at its Capital Markets Day that approximately 46% of exports from its controlling stake in the Tshipi Borwa mine are transported by road, with road transport costs being approximately 37% higher than rail. Logistics account for 43% of the Free On Board (FOB) cost. Jupiter Mines disclosed in its annual report that Tshipi transported approximately 2.6 million tons via rail in the fiscal year ended June 2026, with 828,000 tons exported via the port of Lüderitz in Namibia. Although these two sets of figures differ in scope and timing, they point to the same reality: the cost curve for South African manganese ore is largely a transportation cost curve.
Previous articles by Mineral Cube analyzing Transnet’s capacity bottlenecks have examined their impact on shipping rhythms. The difference this time is that the bottleneck is no longer just “unable to ship out,” but rather “unprofitable to ship out.” The former can be bypassed by trucks; the latter cannot.
03 Breakdown
The Cost Ledger for One Ton of Road-Transported Ore
South32’s own financial data clearly reflects the pressure. In FY2026, South African Manganese achieved an FOB selling price of $3.65 per dry metric tonne degree (dmt.d), lower than the previous year’s $3.71 dmt.d. Unit operating costs rose to $3.22 dmt.d, a 6% increase, driven by a stronger Rand and rising diesel prices. Full-year EBIT was only $2 million, compared to $24 million in the previous year, compressing the operating margin from 13% to 8%.
| Metric | FY2025 | FY2026 | Change |
|---|---|---|---|
| Realized FOB Price (USD/dmt) | 3.71 | 3.65 | -1.6% |
| Unit Operating Cost (USD/dmt) | 3.05 | 3.22 | +5.6% |
| Gross Margin per dmt (USD, estimated) | 0.66 | 0.43 | -35% |
| Revenue (USD million) | 353 | 364 | +3.1% |
| EBIT (USD million) | 24 | 2 | -92% |
| EBITDA (USD million) | 46 | 30 | -35% |
Note: Based on 54.6% equity interest; gross margin per dmt and change rate are MineCube estimates; FY2025 EBITDA is derived inversely from the guidance of “reducing by USD 16–30 million to reach USD 30 million.”
Drilling down further, a scenario analysis using parameters disclosed by Exxaro assumes that logistics account for 43% of FOB costs, with 46% of volume transported by road. Given that road transport is 37% more expensive than rail, we can decompose the average cost of USD 3.22: the FOB cost per dmt via rail is approximately USD 3.02, while via truck it is approximately USD 3.46, representing a premium of about 14% for the latter (estimated). These parameters originate from Tshipi; applying them to South32 is an approximation, but the magnitude is credible.
| Item | Rail Ton | Road Ton | Explanation |
|---|---|---|---|
| FOB Cost (USD/dmt) | ~3.02 | ~3.46 | Decomposed from the average cost of USD 3.22 |
| November CIF Quote (USD/dmt) | 4.35 | 4.35 | South32 South African Semi-Carbonate Lumps |
| Sea Freight Equivalent (USD/dmt) | ~0.92 | ~0.92 | South Africa to Tianjin at USD 34/ton |
| Implied FOB Netback (USD/dmt) | ~3.43 | ~3.43 | CIF minus sea freight |
| Profit Margin per dmt (USD) | ~0.41 | ~-0.03 | Excluding marketing and other expenses |
Data in the table are MineCube estimates: logistics account for 43% of FOB costs; road transport is 37% more expensive than rail; 46% of volume moves by road; sea freight uses the midpoint of USD 33–35/ton, converted based on a 36.9% grade, without deducting moisture content.
The implication of this table is straightforward: reversing from the November quote of USD 4.35, rail-tonned shipments retain a profit margin of approximately USD 0.4 per dmt, while road-tonned shipments are essentially breakeven. Adding marketing expenses, grade discounts, and rand volatility, road-tonned shipments incur losses in many months. South32’s unit cost guidance for FY2027 is approximately USD 3.35/dmt, citing “inflation and rising inland logistics costs,” indicating that the cost curve continues to rise.
Therefore, “produced more, sold less” is not a contradiction, but a rational choice: mine production remains steady, and cargo within rail quotas ships as usual. Cargo exceeding rail quotas, which must rely on truck transport, remains in stockpiles under current prices. What has been cut is not capacity, but the most expensive segment of logistics.
04 Breakdown
Smelters fall first, mines retreat later
The cost pressure in the South African manganese industry chain did not crush mines first, but smelters. Transalloys, South Africa’s last manganese alloy smelter, ceased all furnace operations on July 1, 2026, due to excessively high electricity prices. The plant produced approximately 155,000 tons of manganese alloys annually, supporting about 600 direct jobs and roughly 7,000 downstream jobs. As early as December 2025, it had already shut down three furnaces, reducing capacity utilization to approximately 57%.
The direct cause was the preferential electricity tariff of 62 cents/kWh granted by the South African energy regulator to ferrochrome smelters on May 29, which excluded manganese alloy and ferrosilicon smelters. By late August, Transalloys’ CEO stated that the plant had not laid off workers, maintained environmental compliance, and retained conditions for rapid resumption of operations. However, restarting required approximately ZAR 1.65 billion and necessitated securing a definitive electricity pricing scheme first.
| Time | Event | Implication |
|---|---|---|
| December 2025 | Transalloys shuts three furnaces; capacity utilization drops to ~57% | Electricity pressure begins to materialize |
| May 29, 2026 | Regulator approves preferential electricity tariffs for ferrochrome smelters | Manganese alloys excluded |
| July 1, 2026 | Transalloys shuts all furnaces | Local South African manganese smelting hits zero |
| August 2026 | South32 cites road transport and diesel costs dragging down South African manganese profits | Mine margins drop to 8% |
| September 25, 2026 | Transnet initiates pre-qualification for Ngqura manganese corridor | Logistics solutions earliest post-2030 |
| October 19, 2026 | South32 Q3 report: South African manganese reduces road transport; sales drop 22% | Mines begin contracting marginal shipments |
It should be noted that Transalloys’ shutdown has minimal direct impact on manganese ore export volumes—the ore consumed by its annual capacity of ~155,000 tons of alloys represents a small fraction of the 26 million ton export scale. It is included here because it shares the same logic as South32’s reduction of road transport: two ends of the same coin. Rising energy and logistics costs in South Africa have not been matched by manganese prices, squeezing out the highest-cost segments of the chain one by one. Smelters fall first; mines subsequently contract high-cost logistics.
05 Comparison
Port inventory clearance has not yet been observed
If marginal supply from South Africa is being cleared, China’s ports should reflect this. Import data shows signals, albeit weak ones. Customs data indicates that China imported 1.3864 million tons of manganese ore from South Africa in September, a month-on-month decrease of 18.55%; August imports were 1.702 million tons. However, total manganese ore imports into China from January to September reached 25.9544 million tons, a year-on-year increase of 9.17%, suggesting that South African ore is not lacking overall for the full year.
| Metric | Value |
|---|---|
| Imports from South Africa in 2025 | 17.42 million tons (53% share) |
| Imports from South Africa in August 2026 | 1.702 million tons |
| Imports from South Africa in September 2026 | 1.3864 million tons (MoM -18.55%) |
| Total imports Jan–Sep 2026 | 25.9544 million tons (YoY +9.17%) |
| Total port inventory of manganese ore (week ending Oct 23) | 6.003 million tons |
| Tianjin Port inventory (as of Sep 18) | 5.107 million tons |
Inventory levels show no signs of clearing. For the week ending October 23, national manganese ore port inventories stood at 6.003 million tonnes, approximately 20% higher than the average since October 2023. Tianjin Port, the primary hub for South African ores, held inventories of 5.107 million tonnes in mid-September, accounting for over 80% of the national total. As long as inventory at Tianjin Port remains above 5 million tonnes, a contraction in marginal supply is unlikely to translate immediately into spot price increases.
| Variety | Price | Notes |
|---|---|---|
| Tianjin Port SA Semi-Carbonate | 33.3–34 | RMB/tonne-degree, mid-to-late October |
| Tianjin Port SA High-Grade Ferrous | 28.8–29.5 | RMB/tonne-degree, mid-to-late October |
| Tianjin Port Gabon Lump | 39–39.2 | RMB/tonne-degree, mid-to-late October |
| Tianjin Port South32 Australian Lump | 38.6–39 | RMB/tonne-degree, mid-to-late October |
| South32 SA Semi-Carbonate November Quote | 4.35 | USD/dmt dry degree, CIF, +0.15 MoM |
| Comilog Gabon Lump November Quote | 4.80 | USD/dmt dry degree, CIF, +0.10 MoM |
Note: Jupiter has not yet released its November quotes for China.
A notable detail within the price structure is the divergence between weak port spot prices and rising offshore quotations. South32 increased its November quote for SA semi-carbonate lumps to China by $0.15 to $4.35, while Comilog raised its quote by $0.10; Jupiter has not yet provided its November quote for China. Mines are simultaneously reducing high-cost shipments and raising prices—a classic display of supply-side discipline—while port traders hold high inventories that suppress spot prices. The tug-of-war between these two forces constitutes the current market structure for South African manganese ore.
06 Assessment
Road transport costs are becoming the price floor
Minli Square’s assessment is that road transport tonnage for South African manganese ore is becoming the marginal cost line for global manganese ore: if the offshore CIF price for SA semi-carbonate falls below $4.0–$4.1/dmt dry degree, road transport tonnage will exit en masse, leaving limited downside space below the 33 RMB/tonne-degree level for semi-carbonate at Tianjin Port; however, the 6 million tonnes of port inventory dictates that this clearance will manifest more as “inability to fall further” rather than “immediate rise” over the next one to two quarters.
First, the scale of clearance is substantial. If the approximate 10 million tonnes of road transport exports contract by 10% to 20% due to uneconomic conditions, this represents 1 to 2 million tonnes, equivalent to approximately 6% to 11% of China’s 2025 imports from South Africa (estimated). A marginal contraction does not require a large percentage to alter the pace of inventory accumulation at ports.
| Road Transport Volume Contraction Ratio | Reduced Export Volume (10k tonnes) | Share of China’s 2025 Imports from SA | Equivalent to Monthly Avg Imports from SA |
|---|---|---|---|
| 10% | Approx. 100 | Approx. 5.7% | Approx. 0.7 months |
| 20% | Approx. 200 | Approx. 11.5% | Approx. 1.4 months |
| 30% | Approx. 300 | Approx. 17.2% | Approx. 2.1 months |
Data in the table are scenario estimates by Minli Square, based on an approximate 10 million tonnes base for road transport exports, with China’s 2025 imports from South Africa set at 17.42 million tonnes.
Second, the clearance is elastic, not permanent. The exit of road transport tonnage depends on prices: if CIF quotes return above $4.6 or if Rand and diesel prices fall, cargo currently stored in yards will quickly return to the road. South32 maintaining its production guidance of 2 million wet tonnes unchanged inherently indicates that mines have not cut production; thus, inventory piled in yards acts as a sword hanging over prices, suppressing the height of any rebound.
Third, long-term solutions arrive too late. The Ngqura Manganese Corridor project launched by Transnet in September plans for an initial annual capacity of approximately 12 to 12.6 million tonnes, with a terminal design capacity of approximately 14.7 million tonnes, expandable to 16 million tonnes, at a total investment of 30 to 44 billion ZAR. However, the port authority previously indicated a commissioning timeline of 2030–2031, with prequalification submissions closing only in February 2027. Until then, the ceiling for South African manganese ore exports is determined by rail, while the floor is determined by trucks.
| Project | Parameters |
|---|---|
| Project Name | Ngqura Manganese Export Corridor (Hotazel to Ngqura, approx. 1,100 km) |
| Bidding Progress | Prequalification documents issued Sept 2026, deadline Feb 26, 2027 |
| Total Investment | 30–44 billion ZAR (Terminal approx. 16 billion, Rail 10–20 billion) |
| Capacity Plan | Corridor initial approx. 12–12.6 million tonnes/year; Terminal approx. 14.7 million tonnes, expandable to 16 million tonnes |
| Equity Structure | Transnet holds 51% of operating company, private partners 49%, term approx. 25 years |
| Expected Commissioning | 2030–2031 (Port Authority April estimate), qualification documents do not provide new timeline |
07 Outlook
Three signals to track clearance progress
Tracking Point 1: South32’s Q4 South African Manganese Sales. The deferred shipment will be replenished in the Q4 figures, so a recovery in sales volume next quarter is normal. The key is whether sales remain significantly below production after excluding this single shipment. If the production-sales gap continues to exceed 100,000 wet tonnes, it indicates that cutting road transport is a sustained operational decision rather than a one-off adjustment; also monitor whether unit costs in the FY2027 interim report rise above $3.35/dmt dry degree.
Tracking Point 2: Whether total port inventory can fall below 5.5 million tonnes. Only when total port inventory drops from 6 million tonnes to below 5.5 million tonnes, while the discharge volume of SA ores at Tianjin Port remains above 200,000 tonnes per week, will it indicate that the contraction in marginal South African shipments has truly transmitted to the Chinese spot market. At that point, semi-carbonate at Tianjin Port is expected to reclaim levels above 35 RMB/tonne-degree; if inventory instead continues to accumulate above 6.2 million tonnes, it suggests that clearance is being absorbed by high stocks.
Tracking Point 3: December offshore quotes versus Rand and diesel prices. The December quotes for China from mines such as South32 and Jupiter serve as direct signals of mining pricing discipline. If quotes continue to rise or hold steady while port spots do not follow, it indicates mines prefer to sell less rather than lower prices; conversely, if a weaker Rand and falling diesel reduce road transport costs, road transport tonnage may re-enter the market, pushing the floor lower. Additionally, whether Transalloys secures an electricity tariff solution to resume production serves as a window into observing the South African government’s attitude toward the manganese industry chain.
Rail sets the ceiling, trucks set the floor.
— END —