Platinum Mining In Depth
The Underlying Logic Is Widely Misread

Most people do not know this: in 2021 rhodium alone accounted for 53% of Northam Platinum's total revenue, and 52% of Impala Platinum's. JP Morgan Cazenove published these figures in a research note dated February 25, 2021. Rhodium made up only 5% to 7% of these companies' production ounces. A co-product with negligible output volume was holding up nearly half the industry's revenue. Platinum, the nominal main product, was playing a supporting role in its own mines.

Then rhodium crashed. From nearly $30,000 down to around $4,400 by the end of 2024. An 84% decline. Palladium fell 70% from its peak. Platinum itself declined less than 10% over the same period.

The disaster was not in the platinum price. The disaster was in the basket price. Noah Capital Markets analyst René Hochreiter ran the numbers: stripping out rhodium and palladium revenue, up to half of global PGM production was operating at a loss by the end of 2023. Strip out chrome and copper by-product revenue as well, and the loss-making proportion climbed to two-thirds.

Looking at PGM mining company financials often produces a puzzling reaction: platinum barely moved, so why are these companies suddenly posting massive losses? The answer is in the basket. Without unpacking the basket, the numbers on the income statement are meaningless.

Rhodium went through one of the most violent price swings in modern commodity history.

The trigger for the spike: Anglo American Platinum's converter plant in Rustenburg, known in the industry as the ACP, suffered an unexpected failure in 2020. Amplats declared force majeure. Rhodium's total global annual production is only around one million ounces. The market is small enough that a single smelting facility going offline pushed the price from $6,000 to $29,800. Implats spokesman Johan Theron said in an interview at the time that "the market is always correct and efficient," but acknowledged the ACP failure and the Norilsk incident were black swan events that "clearly elevated volatility and price risk both up and down."

On why rhodium subsequently collapsed, most analysis points to weak demand and the EV threat. That attribution is not wrong. But the most damaging blow came from a direction most analysts were not tracking.

During the period of elevated rhodium prices, Chinese fiberglass manufacturers launched an emergency materials substitution R&D effort. China Jushi and Shandong Fiberglass developed reinforced platinum alloys and ceramic alternatives that drastically reduced or eliminated rhodium content in bushing plates. Phoenix Refining published a detailed breakdown of this process in a late-2024 article titled "The Great Rhodium Unwind." The piece cited Johnson Matthey's analysis attributing the main cause of rhodium's price collapse to this wave of industrial substitution and the inventory liquidation it triggered. Because it was not just new demand that got cut. Old rhodium-containing bushings were decommissioned and recycled for cash. In the financial reports of China Jushi and Shandong Fiberglass, rhodium recovery sales generated hundreds of millions of yuan in profit. Buyers became sellers.

In the autocatalyst space, thrifting happened in parallel. Johnson Matthey and Umicore improved washcoat designs to cut per-vehicle rhodium loadings by up to 50% without breaching emissions regulations. Battery EV penetration in China exceeded 30% of new monthly car sales by 2024.

New production lines exclude rhodium from the bill of materials at the design stage. The EV-driven threat to platinum in autocatalysts follows the same path. One distinction is worth noting here: rhodium substitution was actively driven by industrial users in response to high prices and was completed within a few years. The EV substitution facing platinum is driven by policy and technology iteration, moving more slowly, but equally irreversible.

Chrome rarely comes up in conversation.

SFA Oxford published an analysis in 2024 that specifically broke down how chrome's role in South African PGM miner revenue has changed. Ten years ago, Statistics South Africa's mining data showed platinum at 68.1% of PGM sales revenue. Rhodium and palladium took most of the rest. Chrome was barely a rounding error. Now chrome is the fourth-largest revenue source, behind platinum, palladium, and rhodium.

UG2 ore is naturally rich in chromite. When PGMs are extracted, chromite is separated as a by-product. Chrome ore prices sit at historical highs above $300 per tonne.

A related shift in the supply chain. China has overtaken South Africa as the world's largest ferrochrome producer. South Africa used to smelt ferrochrome domestically and export the finished product, but years of Eskom electricity price increases have gutted the domestic ferrochrome smelting industry. South Africa has pivoted to exporting raw chrome ore to China. SFA Oxford's report noted that the Chinese market now dominates chrome ore pricing. This creates a dependency that is rarely discussed: South African platinum miners' profitability depends not just on PGM prices but also on the Chinese ferrochrome market. Very few sell-side analysts break out the by-product revenue line item and model it separately. In the basket price collapse of 2023 and 2024, chrome revenue may have been the reason certain mines were not shut down.

Approximately 90% of the world's known PGM reserves are locked in South Africa's Bushveld Igneous Complex. The largest layered igneous intrusion on Earth. Approximately 66,000 square kilometers. Formed by a magmatic intrusion event roughly two billion years ago. During cooling, PGMs were concentrated into specific thin layers due to their high affinity for sulfide melts.

For context on the concentration: the top five copper-producing countries together account for roughly 60% of global output. For iron ore, about 75%. For platinum, South Africa alone accounts for roughly 70%. No comparable concentration exists among major industrial metals.

Three reefs.

The Merensky Reef. Discovered in 1924. Based on cross-referenced descriptions from Wikipedia and multiple geological papers, this reef is approximately 46 centimeters thick, bounded by thin chromitite stringers on both sides. Average PGM grade around 10 ppm, hosted in pyrrhotite, pentlandite, and pyrite, with copper, nickel, and gold as by-products. The debate over how it formed has been running for almost a century. The traditional view argues for gravity settling: crystals sinking through magma to the chamber floor. A paper published in Scientific Reports in 2019 (Latypov et al.) reported the discovery of the reef occurring on subvertical to overturned margins of depressions in a temporary chamber floor. Gravity settling is physically impossible at these orientations. The paper concluded that the reef formed at least partly through in situ crystallization: immiscible sulfide droplets sequestering noble metals directly on the chamber floor and walls in a strongly convecting environment. This has implications for exploration, because it expands the range of geological configurations where mineralization might occur.

The reef contains a structure called a pothole. Roughly circular depressions where floor rocks have been eroded by magma, causing the reef to suddenly thicken or vanish entirely. The destructive impact on mine planning is that a continuous mining machine can move from rich ore into completely barren rock within a few meters. Potholes force mines to maintain geological drilling densities far higher than for other commodities. This expenditure almost never appears as a separate line item in external analysts' cost models.

The UG2 Chromitite Reef. SFA Oxford's data shows it currently provides over two-thirds of South Africa's primary platinum supply. The Merensky's contribution has fallen to around 20%. In the early 2000s a batch of mines started up exclusively on UG2, and since then many formerly Merensky-only operations have added UG2 to their mining plans. UG2 PGM reserves may be nearly twice those of the Merensky. The tradeoff is metallurgical trouble: high chromite content causes refractory spinel deposits to form in electric arc furnaces. A 2018 review published by IntechOpen (Extraction of Platinum Group Metals) provided specific temperature parameters: typical matte operating temperatures of 1,380 to 1,600°C, slag temperatures of 1,500 to 1,680°C. The chrome problem pushes converter matte temperatures above 1,355°C, damaging refractory linings. The paper also mentioned deep electrode immersion operation and graphite block linings that form protective frozen skulls at the hot face as current mitigation approaches. Not a problem that engineering optimization can eliminate.

The Platreef sits on the northern limb of the Bushveld. Tens of meters thick, lower grade, suited to large-scale mechanized mining. Beyond the Bushveld: Russia's Norilsk (platinum is a by-product of nickel-copper mining), Zimbabwe's Great Dyke, and Canada's Sudbury. Together these account for less than 30% of global output.

BullionByPost published an accessible figure: processing 7 to 12 tonnes of ore through the entire chain yields one troy ounce of pure platinum. The process takes eight weeks to six months.

Four stages.

Froth flotation. Roughly 30:1 upgrading. Recovery rate approximately 85%. About 15% of PGMs are permanently lost in tailings at this step, the single largest point of loss in the entire value chain. Crushed ore is ground to fine powder, mixed with water and xanthate-type collectors. In flotation cells, compressed air is blown through the slurry, sulfide mineral particles attach to bubbles and float to the surface to be skimmed off as concentrate.

Smelting. Roughly 10:1 upgrading. Electric arc furnaces, 1,350 to 1,600°C. Material separates into sulfide matte (containing nickel, copper, and PGMs) and silicate slag.

Converting and matte treatment. Roughly 3:1 upgrading.

Precious metal refining. Roughly 200:1 upgrading. South Africa's three major producers each use different technical routes: Lonmin uses precipitation, Anglo Platinum uses solvent extraction, Impala Platinum uses ion exchange. The six PGMs have extremely similar chemical properties. Separation requires repeated cycles of dissolution, precipitation, and extraction. The entire refining cycle can last months.

Pipeline lock-up. During the five-to-six-month processing cycle, PGMs worth billions of rand are stuck in the concentrator, the smelter, the refinery. They cannot be sold or delivered. They just sit there, locked in, waiting to clear the pipeline. The same technical document notes that the value of precious metals permanently locked inside refining equipment sometimes exceeds the capital cost of the equipment itself. That is why Anglo American Platinum invested heavily in the slow-cool process. The specifics: converter matte is cast into 30-tonne ingots in refractory-lined moulds and slowly cooled for five days. PGMs concentrate in a ferromagnetic nickel-copper-iron alloy that can be recovered by magnetic separation, skipping the BMR and going directly to the PMR. What gets saved is not metallurgical cost. It is working capital.

The Valterra Platinum example makes more sense in this context. Spun out of Anglo American and independently listed. Nedbank's Van Graan mentioned in an October 2025 research note that Valterra increased capital investment in its Rustenburg processing facilities from just 500 million rand in 2016 to approximately 6 billion rand in 2024. His assessment was that these downstream assets constitute "a key strategic advantage which should support its premium valuation multiples relative to peers." Miningmx quoted this in an October 2025 article. Share price up 72% since IPO. Valterra can produce refined metal from less concentrate. Its competitors are still stacking up concentrate waiting for smelting slots. When the basket price is good this gap is not very visible. When the basket price is bad it is the difference between positive cash flow and scrambling for working capital.

South Africa's smelting and refining pipeline is constrained by a shortage of BMR capacity. Smaller miners without their own smelting and refining capability have no choice but to sell flotation concentrate to major producers for toll processing. In platinum mining, the strategic value of downstream processing capacity may rival that of the mining right itself. Platinum Group Metals signed an MOU with Saudi Arabia's Ajlan & Bros to study the construction of an independent smelter and BMR in Saudi Arabia. The SEC's public filings contain a record of this (a Cooperation Agreement announcement dated December 20, 2023). If this path succeeds, the global PGM refining map would see its most significant change since the 1920s. It is still at the feasibility study stage.

The number of operational PGM shafts in South Africa fell from 81 in 2008 to 53 in 2025. Nedbank Securities data. A 35% reduction. Nedbank estimates a further decline to approximately 47 by 2030.

Output per shaft rose from 95,000 ounces in 2008 to approximately 126,000 ounces. The industry shifted from many small shallow shafts to fewer large deep shafts. Total production kept declining. Johnson Matthey's historical data shows South African platinum output peaked at approximately 5.3 million ounces in 2006. It has trended downward ever since. Two consecutive years below 4 million ounces, the threshold commonly used to gauge whether the industry is "healthy."

Northam CEO Paul Dunne said several things at the annual results presentation on August 30, 2024 (Reuters published a report the same day). He estimated that industry platinum output would decline by roughly 500,000 ounces every five years. He said: "We haven't replaced that asset base and the asset base is a depleting asset, all mines are depleting assets from day one." He expected South African output to fall about 10% over the next five years to around 3.5 million ounces. Impala Platinum CEO Nico Muller told Reuters around the same time that building new mines in South Africa was "highly improbable." Dunne added that the "depressed pricing environment" could persist for as long as two years. As it turned out, the platinum price did rally in 2025, but Northam's fiscal year ending June 2025 came in with headline earnings per share of 3.81 rand, down from 4.45 rand the prior year. This was achieved on record sales volumes. Dunne's language at that results presentation shifted to: prices are "still not yet at levels that will support sustainable mining across the industry and certainly not the much-needed development of new operations." Record sales volume, declining profit. Costs ate up all the revenue gains.

Industry analysts' estimates, compiled in a Discovery Alert report from September 2025: existing operations need the platinum price to stabilize consistently above $1,500 to $1,700 per ounce for sustainable operation. New mine development requires $1,800 to $2,000. The current price is around $979. Well below both thresholds. Many mines are still running on the subsidy of chrome and residual rhodium and palladium revenue in the basket price, plus the cushion from rand depreciation on dollar-denominated revenue. If any of these crutches gives way, whether chrome prices retreat or the rand strengthens, more shafts will fall below breakeven.

African Rainbow Minerals and the Bokoni mine. Acquired in 2022 for 3.5 billion rand. Then found to require 15 billion rand in capital expenditure and a seven-year development timeline. Forced to take a 2.2 billion rand write-down. Phoenix Refining cited this case in a November 2025 market analysis. Acquisition price 3.5 billion. Development cost more than four times that. Write-down before development was even complete.

Global PGM refined production fell 8% year-on-year in the first half of 2025, Metals Focus data. Hochreiter projected South African platinum full-year output down 10% year-on-year to approximately 3.65 million ounces. Flooding, processing bottlenecks, power disruptions, and depletion of work-in-process inventory all converging.

The platinum market has recorded supply deficits for three consecutive years. WPIC's Q3 2025 Platinum Quarterly report provided the specific numbers: above ground stocks depleted by 42% (approximately 2,341 koz) over three years, with less than five months of demand cover remaining. 2025 total supply approximately 7,129 koz, down 2% year-on-year. Total demand approximately 7,821 koz, down 5% year-on-year, with industrial demand declining 22% as the glass capacity expansion cycle hit a trough.

Everyone reports the supply-demand figures. What is more telling is two technical indicators.

Platinum's lease rate. The fee that physical holders of platinum charge counterparties who need to borrow the metal temporarily. WPIC's report specifically flagged this signal: rates have been persistently elevated, indicating that physical holders expect to need to retain their metal to fulfill delivery obligations and are unwilling to lend. The London OTC market simultaneously exhibited backwardation: spot prices above forward prices. In commodity markets this is typically read as a sign of physical tightness.

WPIC's report contained a carefully worded observation: "sustained elevated lease rates and strong backwardation highlight that balancing flows from above ground stocks to fill the 2025 deficit have been insufficient to ease tight market conditions." WPIC's view was that a "substantial market surplus" might be needed to alleviate tightness going into 2026.

There is a difference between platinum and gold in physical terms that is easy to overlook. Gold, once mined, overwhelmingly persists on the surface as bars, coins, and jewelry. Total above-ground gold stocks exceed 200,000 tonnes. It can be melted and recast indefinitely. Platinum, once mined, is mostly consumed by industrial processes. It goes into autocatalysts. Into glass furnace bushings. Into chemical reactors. A significant portion is permanently lost in dispersed form or requires dedicated recycling processes to recover. When industrial consumers start having difficulty sourcing material, lease rates and spot premiums send the signal before the price itself moves. Right now the signals are there.

Van Graan said in an interview with S&P Global Platts: "There is no such thing as quick ounces. When talking about rhodium ounces, at best it is 12 months away when talking about a meaningful increase in rhodium output." Standard Chartered's precious metals analyst Suki Cooper expressed a similar judgment. Her exact words in an interview with Platts: "Demand can respond over a matter of weeks, mine supply takes years." Once a supply-demand gap forms it persists in the market. It does not clear quickly the way it might in oil markets with their elastic supply mechanisms.

Ruthenium. The least conspicuous member of the PGM family. Currently a small fraction of the basket price.

Hochreiter believes ruthenium prices reaching $1,000 per ounce or more are "not far away." He expects ruthenium's price trajectory could follow platinum's over the coming months to years. The driving force would be ruthenium-based materials finding application in next-generation data storage technology and AI infrastructure. This is currently a minority view. Ruthenium's applications in AI remain at the R&D and early commercialization stage, which is a different situation from rhodium's decades-proven rigid demand in autocatalysts.

Sibanye-Stillwater's head of corporate affairs James Wellsted disclosed a data point in a Miningmx report from August 2025: a ruthenium price move from $675 to $800 would add approximately $33.1 million to annual revenue. Sylvania Platinum CEO Jaco Prinsloo offered another perspective in the same article: a 10% rise in rhodium increases Sylvania's total revenue by 2% to 3%, because approximately 12% of the company's output is rhodium.

If AI-driven ruthenium demand materializes, it could repeat the role rhodium played in 2020 and 2021. Ruthenium's annual production is even smaller than rhodium's. The lesson from the rhodium episode is that a small market hit by concentrated demand sees price movements that far exceed any model's forecast range. Of course the lesson also includes the other half: once the demand shock fades or gets substituted away, the price decline is equally violent.

Autocatalysts account for approximately 45% of total platinum demand. Tighter emissions standards are increasing per-vehicle platinum loadings. Price-driven substitution of palladium with platinum is underway. Battery EV penetration in China exceeded 30% of new monthly car sales in 2024. Tightening emissions standards push up per-vehicle loadings while electrification shrinks the total number of vehicles carrying catalytic converters. The net effect depends on relative speeds. Neither side has gained a decisive advantage so far. SFA Oxford's analysis pointed out that roughly two-thirds of South African ore comes from UG2, and UG2 has greater exposure to palladium and rhodium prices than the Merensky Reef, meaning palladium and rhodium price declines hit South African miners' revenue harder than they would hit a Merensky-dominated operation.

Jewelry accounts for approximately 25%. Chinese platinum bar investment demand grew roughly sixfold between 2019 and 2023, from 0.87 tonnes to 5 tonnes, per Straits Research data. With gold at historical highs, price-sensitive consumers have begun shifting toward platinum. Chinese platinum jewelry demand was approximately 453,000 ounces in 2024.

Industrial applications account for approximately 23%, spanning glass manufacturing, chemical catalysis, and electronics. WPIC data showed industrial demand down 22% year-on-year in 2025, primarily because the glass capacity expansion cycle was at a trough, with glass-related platinum demand plunging 74% year-on-year. A rebound is expected in 2026.

Hydrogen. Proton exchange membrane fuel cells require approximately 30 to 60 grams of platinum per vehicle. Traditional autocatalysts use 3 to 7 grams. Fuel cell technology is working to reduce platinum loadings. The competition between alkaline electrolyzers and PEM electrolyzers remains unresolved, and the two pathways differ by orders of magnitude in their platinum demand implications. In all discussions of platinum demand, hydrogen is the segment with the most optimistic assumptions and the highest uncertainty.

Approximately 70% of global platinum supply depends on South Africa.

Eskom's power grid has been chronically unstable. The 2012 Marikana incident. The 2014 strike that lasted five months. In November 2023 the conveyance system at Impala Rustenburg shaft malfunctioned. Mining Technology described the incident in detail in a May 2024 article: the system carrying workers to the surface from approximately 1 kilometer underground began a rapid uncontrolled descent. Eleven workers died. Seventy-five were injured. Implats halted operations temporarily. Its Johannesburg share price dropped 8% on the day. South African platinum and gold mines reported 49 fatalities in 2023, the lowest number on record. Forty-nine deaths is "the lowest."

The PGM industry directly employs over 181,000 workers, per 2023 data from the Minerals Council of South Africa. Sibanye-Stillwater closed multiple shafts over the past 18 months, including Beatrix 4, Kloof 4, Kroondal Simunye, and Marikana 4B. South Africa's unemployment rate exceeds 30%. The social shockwave from shaft closures far exceeds the write-down figures in financial reports.

Zimbabwe. GlobalData projected in a September 2025 report that platinum production would grow at a 2.9% CAGR to 616 koz by 2030. The Mupani and Karo projects would together add 64 koz of annual capacity. Still small relative to South Africa.

Russian platinum output in 2025 was estimated at a slight decline to approximately 742,800 ounces. Mining Technology reported in July 2025 that this was due to the closure of the first stage at the Oktyabrsky mine. Western sanctions have not directly targeted PGMs. Indirect shifts in trade flows are accumulating.

The EU has classified PGMs as critical raw materials. The US Commerce Department's Section 232 investigation and China's revocation of its platinum import tax rebate could create tiered pricing structures between the US, China, and the rest of the world. WPIC's Q3 2025 report mentioned this possibility. If tiered pricing forms, platinum's global arbitrage opportunities and circulation efficiency would both change. South African miners, who depend on international markets, would be most affected.

Recycling platinum requires 10% to 20% of the energy needed for primary mining. Karat24 cited this figure in an early 2026 review. Recycled PGMs already account for nearly 40% of market supply in advanced economies. WPIC projects double-digit growth in secondary supply for 2026, driven by higher prices.

Growth in recycling volumes simultaneously creates a hidden ceiling on the platinum price. When prices rise enough to activate recycling of dormant autocatalyst inventories and industrial scrap, the influx of secondary supply suppresses further upside. Rhodium already demonstrated this mechanism. Not only were Chinese fiberglass bushings torn out and sold, but autocatalyst recycling volumes globally also jumped during the high-price period. Platinum's recycling elasticity is not as extreme as rhodium's. Platinum recycling sources are more dispersed, per-unit platinum content in recyclable items is lower, and the recycling logistics chain is longer. But the direction is the same. When modeling upside scenarios for the platinum price, the price-response function of recycling supply is a binding constraint.

Global annual platinum production is approximately 170 tonnes, Straits Research's 2025 data. Gold is approximately 3,600 tonnes. Before 2008, platinum's price consistently exceeded gold's. Currently around $979 per ounce. Gold above $3,300. Platinum sits roughly $800 below its inflation-adjusted all-time peak. WPIC specifically highlighted this spread at the end of its Q3 2025 report.

NAI 500 listed three indicators in a February 2026 analysis of Northam Platinum, arguing they hold more forward-looking value for mining company investors than macro forecasts: net cash or leverage trend, committed project capital expenditure, and working capital tied up in the concentrate-to-refined-metal pipeline. SBG Securities' Adrian Hammond favors Northam and Valterra, with the rationale quoted in an October 2025 Miningmx report: both companies "present growth, cost and capital discipline while remaining countercyclical." Citi noted Northam has tripled production organically over the past decade, with a further 6% increase on the cards over the next four to five years. The NAI 500 article also flagged a detail: if rhodium and iridium refining hits bottlenecks, it slows cash conversion across the entire pipeline. This kind of bottleneck is especially likely to occur during months when Eskom load-shedding coincides with smelter maintenance, causing uneven distribution of sales revenue across reporting periods.

The variables are concentrated on the demand side. The speed of EV encroachment on autocatalysts. The pace at which hydrogen materializes. Whether ruthenium's AI application prospects become reality. How Chinese consumers allocate their budgets between gold and platinum. These variables operate independently, with little correlation between them, each on its own timeline and driven by its own forces. Whether platinum continues to be overlooked or gets repriced over the next decade depends on which of these variables moves first.