|
Part I
The Mechanism
|
Manganese.
The sell-side doesn't cover it. Bloomberg calls it "forgotten." CNBC has literally never, in living memory, led a segment with the word "manganese." Its ore sells for less per tonne than a decent espresso machine. If you stopped ten commodity analysts in a hallway and asked them to name the battery-grade manganese processing leader by country, eight of them would guess Australia. They'd be wrong by about 95 percentage points.
Here's the part nobody on the terminal screens is watching: China controls roughly 95% of global high-purity manganese sulphate monohydrate production — the refined form that goes into every NMC, LMFP, and LMNO battery cathode rolling off the line. That's the same concentration level that triggered export controls on gallium, germanium, and rare earths. And manganese is already on Beijing's 2023 Critical Minerals List. The gun is loaded. It just hasn't been fired.
Meanwhile, CATL and BYD are scaling LMFP — lithium manganese iron phosphate — as the next-generation cathode chemistry. It's cheaper than NMC, denser than LFP, and it eats more manganese per kilowatt-hour than anything currently in mass production. Benchmark Minerals projects battery-grade manganese sulfate demand to increase 5.9x by 2030. The market is building a cathedral out of a material one country refines.
It's like pouring a concrete foundation and then discovering the only cement plant is owned by the guy you're in a trade war with.
I’m about to get into this Cybercab… and it’s going to take me to what Elon Musk called…
“The most epic AI project ever.”
Because this is all part of Elon Musk’s $320 trillion masterplan.
|
Part II
The Diagram
|
Story off. Numbers on.
South Africa's Kalahari Basin holds 80% of the world's known manganese ore reserves. The country produces around 7.6 million tonnes annually — about 38% of global output. Gabon adds another 7.1 million through Eramet Comilog. Australia chips in 3.0 million via South32.
But ore is not the bottleneck. Ore is cheap. Eramet's October offer for 44.5% Mn lump landed at $4.70 per dry metric tonne unit CIF China. That's the price of a latte per unit of the fifth most abundant metal in the earth's crust.
The bottleneck is refining. Converting that ore into HPMSM requires specific chemical processing, stringent purity controls, and factory-scale infrastructure that exists almost nowhere outside China. Battery-grade MnSO₄ trades around 7,000–7,650 yuan per tonne inside China. Outside China, there is barely a commercial-scale facility to quote a price against.
China imported 25.95 million tonnes of manganese ore in January through September 2026 — up 9.3% year-on-year. The ore is flowing in. But what flows out as refined product is entirely at Beijing's discretion. That's not a supply chain. That's a permission slip.
And even the ore side has problems. South Africa's Transnet — the state-owned rail operator — is seeking a private partner for a R44 billion ($2.7 billion) manganese export corridor because the existing rail infrastructure is failing. The proposed system would handle about 12 million tonnes initially, potentially rising to 16 million. Rail rehabilitation alone could require R20 billion.
Jupiter Mines' Tshipi operation has already rerouted 828,000 tonnes through Namibia's Port of Lüderitz — a workaround that tells you everything about the state of the primary corridor. When miners start shipping through neighboring countries to avoid your railway, your railway is not a railway. It's an obstacle.
|
Part III
The Weak Link
|
China hasn't pulled the trigger on manganese export controls yet. That's the part that should keep you awake.
Manganese sits on Beijing's 2023 Critical Minerals List alongside gallium, germanium, rare earths, and tungsten. Every single one of those has been weaponized with export licensing, quotas, or outright bans in the past three years. Manganese is the last major card in the deck. And unlike those others, the market isn't even slightly positioned for it — because the market doesn't think manganese matters.
The mainstream read is that manganese is abundant and cheap. That's true — at the ore level. It's the same logic that said graphite was in oversupply while 97% of processing sat inside China. The oversupply exists inside China's borders. Outside those borders, battery-grade manganese is functionally unavailable at scale.
Euro Manganese's Chvaletice project in the Czech Republic — the West's most advanced non-Chinese HPMSM facility — is still in demonstration phase. Its May 2026 PEA estimates operating costs at $4.86 per kilogram of manganese equivalent, well above Chinese competitors. South32 is exploring options. The Manganese Metal Company in Mbombela commissioned a 6,000-tonne-per-year commercial battery-grade sulfate plant in 2026, but that's a rounding error against China's output. The non-Chinese pipeline for HPMSM is still dwarfed by China's installed capacity. I've seen this movie before with graphite. The ending wasn't good for the buyers who assumed the supply would always be there.
Ninety percent of manganese demand is still steel and ferro-alloy — the boring stuff. But the marginal tonne of new demand is all batteries. And marginal demand is what sets price when supply is constrained at the processing stage. The steel guys won't notice the squeeze. The battery guys will.
|
Part IV
The Chain Reaction
|
The trigger isn't hard to imagine. Beijing has escalated export controls methodically — gallium and germanium in 2023, antimony in 2024, rare earth controls escalated in 2025. Each time, the pattern is identical: announce controls, suspend briefly for diplomatic leverage, then ratchet tighter. Manganese is next in the sequence. The question is when, not if.
If HPMSM exports get licensed, the immediate casualty is every non-Chinese cathode maker sourcing battery-grade manganese sulfate — which is virtually all of them, because there's nowhere else to buy it at scale. CATL and BYD, operating domestically, are insulated. Their Western competitors are not. This is not an accident. It's an industrial strategy that's been building for a decade.
I watched this exact sequence play out with gallium in late 2023. The market shrugged. Then exports went to zero overnight. GaN semiconductor buyers spent six months scrambling for alternative supply that didn't exist. Manganese is a bigger market, harder to substitute, and the alternative infrastructure is even thinner.
Where does capital go? Not into the ore miners — ore is not the chokepoint. South32 and Eramet will keep shipping rock to China at $4.70 per unit regardless. The edge, if there is one, sits with the handful of companies building non-Chinese refining capacity: Euro Manganese's Chvaletice project, South32's downstream ambitions, any junior that can credibly produce battery-grade sulfate outside Beijing's jurisdiction. They are years from commercial production. But in a world where the last card gets played, "years away" starts looking like "the only game left."
The broader play is cathode-agnostic EV supply chain companies that can pivot chemistry or source from multiple jurisdictions. But the pure exposure — the leveraged bet on a manganese processing squeeze — is razor-thin. There are maybe three public names globally with credible non-Chinese HPMSM projects. That's not a sector. That's a rounding error.
Financial layer says manganese is cheap and abundant. Physical layer says the refined product is a monopoly with a known history of weaponization. In my experience, when the financial layer discovers what the physical layer already knows, the repricing is not gentle. The only question is whether Beijing plays the card before the West builds the refinery. Right now, the refinery is a PowerPoint deck. The card is in Beijing's hand.
Sources: Benchmark Minerals, Shanghai Metals Market, Eramet Comilog, South32, Transnet, Fastmarkets, E Source, MOFCOM, Jupiter Mines, Euro Manganese
