|
Part I
The Mechanism
|
Manganese.
The sell-side calls it "abundant." The commodity strategists skip past it to write their eleventh lithium note of the quarter. Bloomberg's metals page barely tracks it. And they're all correct about one thing: there is plenty of manganese ore in the ground. South Africa alone sits on 40% of global reserves. You can find the stuff on five continents.
But ore isn't what goes into a battery cathode. What goes into a battery cathode is high-purity manganese sulfate monohydrate — HPMSM — refined to 99.95% purity through a process complex enough that one country controls virtually the entire global supply. That country is China. The share, depending on whose estimate you trust, runs between 90% and 96%.
The mainstream still models manganese as a steel input — and for 90% of current tonnage, it is. But the demand vector that's about to matter is batteries. BYD launched Blade 2.0 in March with an LMFP cathode. CATL and Gotion are scaling the same chemistry. Benchmark Mineral Intelligence estimates BYD's LMFP shift alone could increase 2026 battery-grade manganese demand by 7%. And manganese sulfate demand overall is forecast to grow ninefold by 2030.
The analysts who ignored manganese because it was cheap are about to discover that cheap is a function of nobody needing the refined product. That's changing. And the refining capacity to meet it doesn't exist outside China.
|
Part II
The Diagram
|
Strip the narrative. Here's how the manganese machine actually works.
Three countries produce roughly two-thirds of global manganese ore: South Africa (~7.2 million metric tons/year, roughly 40% of exports), Gabon (~7 million MT), and Australia (~1.6 million MT). China also mines around 0.7 million MT, but it consumes more than 52% of global ore — it is a net importer by a wide margin. All roads lead to Chinese ports and Chinese refineries.
Now layer in the disruptions. South32's GEMCO mine on Groote Eylandt — one of the world's largest high-grade manganese operations — just cut FY2026 production guidance by 6% to 3.0 million wet metric tons. The reason: Tropical Cyclone Narelle. This is the second cyclone to hit GEMCO in two years. Cyclone Megan shut the mine for four months in 2024 and doubled manganese ore prices. Exports only resumed in May 2025. The market barely had time to recover before Narelle arrived.
High-grade ore was assessed at $6.02/dmtu CIF Tianjin in mid-August after Eramet Comilog lowered the Gabonese lump price for September shipments. Volatility is the signal. The all-time high was $7.96/dmtu back in November 2016. The margin between current prices and that record is thinner than the market appreciates.
The USGS says it plainly: manganese has no satisfactory substitute in its major applications. None. Not in steel. Not in batteries. The machine has no bypass valve.
Every node above the refinery is diversified. Every node below it is captive. That's the diagram. Stare at it long enough and one question emerges: what happens when the refinery node tightens while demand accelerates underneath it?
|
Part III
The Weak Link
|
Three fractures are propagating through this system simultaneously. The market is pricing in zero of them.
First: Gabon. On May 30, President Oligui Nguema announced a full ban on raw manganese ore exports, effective January 1, 2029. Gabon is the world's second-largest producer. Eramet Comilog operates the world's largest manganese mine there and produced roughly 7 million tonnes in 2025. The government wants domestic processing — alloys, battery-grade oxide — before any ore crosses a border. They've signed an MOU targeting 700,000 tonnes per year of local processing by 2031. That's a fraction of what currently ships out. If the ban holds, the seaborne market loses its second-largest source of high-grade ore.
I've heard this kind of announcement before from resource-rich nations. Sometimes it sticks, sometimes it softens. But Gabon isn't bluffing in a vacuum — this is the same resource nationalism playbook running through the DRC on cobalt, Zimbabwe on lithium and PGMs, and Indonesia on nickel. The pattern is clear enough that you'd be reckless to bet against it.
Second: South Africa's logistics. Transnet issued a $2.7 billion Request for Qualifications on September 23 — five days ago — for the Ngqura Manganese Export Corridor, a 25-year public-private partnership to rebuild the rail link from Hotazel in the Northern Cape to Eastern Cape ports. The fact that this project even exists tells you how broken the system is. Exxaro disclosed that logistics account for 43% of free-on-board export costs. Road haulage — which picks up the slack when Transnet's rail fails — costs 37% more than rail. The world's largest manganese producer is hemorrhaging margin before its ore reaches a ship.
Third — and this is the one that keeps me up — the demand side. LMFP is not a lab curiosity anymore. BYD shipped Blade 2.0 with a lithium manganese iron phosphate cathode in March. It raises the voltage platform from 3.2V to 3.8V, delivering 15–20% more energy density while keeping LFP's safety profile. Gotion is ramping LMFP production alongside its existing chemistries. Battery-grade manganese sulfate demand is forecast to grow ninefold by 2030 and twentynine-fold by 2050 — the fastest growth rate of any key battery metal.
And here's the part that makes the machines look foolish: Chinese HPMSM facilities averaged only 45% capacity utilization in 2024, down from a historical 60–65%. Several plants halted operations entirely. The sell-side looked at those numbers and concluded "oversupply." What they actually showed was a market in transition — old steel-grade demand stalling while battery-grade demand hadn't yet arrived at scale. The capacity exists on paper. The qualified, battery-grade capacity running at commercial scale does not.
The U.S. hasn't mined manganese ore since 1970. It is 100% net import reliant. Europe is building gigafactories that need HPMSM it cannot produce domestically. And the two countries that mine the most ore — South Africa and Gabon — are each breaking in different ways at the same time. One can't move its ore to port. The other is about to stop exporting it altogether.
|
Part IV
The Chain Reaction
|
The sequence here is slower than a copper squeeze but structurally harder to reverse. Copper at least has an active futures market full of speculators who can be squeezed out overnight. Manganese refining capacity takes years to build, qualify, and ramp. There's no panic-bid exit valve.
The critical variable is timing. Fastmarkets projects the battery-grade sulfate market flips to deficit by 2027–2028 even if all announced non-Chinese projects hit their targets. CPM Group models the high-purity manganese deficit reaching 475,000 tonnes manganese-equivalent by 2031. If LMFP adoption accelerates faster than the base case — and BYD's Blade 2.0 suggests it will — the deficit pulls forward.
When Cyclone Megan shut GEMCO in 2024, high-grade ore prices doubled in weeks. That was a single mine disruption. What I'm describing here is three structural forces — Gabon withdrawal, South African logistics decay, and a chemistry-driven demand inflection — converging on a refining bottleneck that has not been expanded outside one country. That doesn't double prices. It reprices the entire cost curve.
Where does capital go? The edge — if it exists — is narrow and specific. South32 is building the first new U.S. manganese mine in decades on the Fast-41 permitting track. GM loaned $85 million to Element 25 for a domestic HPMSM processing plant. Euro Manganese, Giyani Metals, and Manganese Metal Company in South Africa are all racing to commission non-Chinese refining capacity. These are small names, most of them pre-revenue. That's the point. The re-rating hasn't started because the market hasn't priced in the chokepoint.
Broad commodity ETFs won't capture this. They're weighted toward steel-grade manganese miners whose revenue rises with construction cycles, not battery chemistry shifts. The levered play is midstream — the refinery link between ore and cathode. The one link that barely exists outside China.
Manganese was abundant, cheap, and boring for a long time. It won't stay all three. The ore is still abundant. The refining never was. And BYD just made sure it won't be boring much longer. In my experience, the market figures this out about eighteen months after the loading docks do. The loading docks are already telling you.
