Platinum Group Metals — A Complete Market Guide (2026)

Data as of 6 July 2026. Prices are quoted as multi-year and full-year averages, not a single day’s snapshot, so this report stays useful over time. Reserves, production splits, balances and historical series are estimates from agency data, rounded for clarity. This report is for information only and was prepared with AI assistance — see the disclaimer at the end.

The platinum group metals market is the strangest corner of commodities: six precious metals rarer than gold, mined at scale in exactly two places on Earth, and consumed mostly by a single technology — the catalytic converter — whose days are numbered. That collision of scarcity and obsolescence makes the platinum group metals market the great cycle machine of the precious complex: palladium delivered a six-year supercycle and then lost two-thirds of its value; rhodium went from $700 to nearly $30,000 an ounce and back; and platinum, after a decade in the wilderness, staged 2025’s most spectacular commodity rally — four consecutive supply deficits driving it to all-time highs. This report is the free, big-picture primer on how the PGM market actually works — the six metals and their baskets, the South African-Russian duopoly, the autocatalyst engine and its electric twilight, and what actually drives returns. For the company-level data behind the charts — every producer screened by production, resources and cost — go to Metal Pilot.

TL;DR & Key Takeaways

• What they are: six sister metals — platinum, palladium, rhodium plus the minors iridium, ruthenium and osmium — always mined together, prized for catalytic power and extreme durability. The dominant use is autocatalysts (~40% of platinum, ~80% of palladium, ~85% of rhodium demand).
• Market structure: the most concentrated primary supply in metals: South Africa’s Bushveld Complex (~70% of mined platinum, ~83% of world reserves) and Russia’s Norilsk (top palladium producer, as a by-product of nickel ) — with recycling of scrapped catalytic converters the third pillar at roughly a quarter of supply.
• Demand story: a race between decline and discovery. Combustion-engine catalysts fade with electrification, but hybrids (PGM-loaded) are booming, platinum is substituting into palladium’s uses, jewellery and industry are steady, and hydrogen (PEM electrolysers and fuel cells need platinum and iridium) is the long-dated option.
• Price regime: violent, metal-specific cycles around a precious-metal core: palladium’s supercycle peaked at $3,440/oz in March 2022 and collapsed to a $995 average in 2024; platinum averaged $961 in 2024, then — on a third straight deficit and vanishing stocks — rallied 84% in 2025 (**$1,200 average**) to a record $2,414 in December 2025.
• Biggest swing factor: South African supply attrition (deep, power-starved, loss-making mines closing shafts) against the pace of the auto-drivetrain transition — with above-ground stocks, now under four months of platinum demand, the shrinking buffer between them.

Numbers to remember (PGMs at a glance)

Figure 1. PGMs at a glance

Figure data: USGS Mineral Commodity Summaries 2026 and WPIC Platinum Quarterly ; see Sections 1.3–2.7.

Why it matters now: the PGM market has flipped from priced-for-obsolescence to physically short. Years of loss-making prices closed South African shafts and idled America’s only PGM mine; the EV transition slowed just as hybrids surged; and by 2025 the platinum market was running its third deep deficit with above-ground stocks approaching exhaustion — the setup behind the rally that carried it to records. For investors, PGMs are where precious-metal safety, deep-value mining equities and a genuine structural squeeze intersect. The big-picture case is below.

How to read this guide: new to PGMs? Read straight through from Section 1. Here for the investment case? Jump to Section 4 — how to own them, how to screen the miners, and what drives the cycle. Every technical term is defined in the Vocabulary (Section 7).

1. PGMs & the market basics

1.1 What PGMs are — physical basics & quality

The platinum group metals — platinum (Pt), palladium (Pd), rhodium (Rh), iridium (Ir), ruthenium (Ru) and osmium (Os) — are six chemically similar, silvery-white metals that sit together in the periodic table and, crucially, together in the ground: every PGM deposit yields a basket of them in fixed geological ratios, the same joint-product economics that rule rare earths . They are genuinely rare (all six together are mined at under 400 tonnes a year — gold’s output is ~9× larger) and genuinely useful: their surfaces catalyse chemical reactions without being consumed, they resist corrosion at extreme temperatures, and they are dense, ductile and stable. That catalytic gift created the modern market: since the 1970s, the three-way catalytic converter — a honeycomb coated with Pt, Pd and Rh that scrubs exhaust gases — has been the demand engine, taking the majority of all three major metals.

Each metal has its own market personality, defined here once:

• Platinum — the diversified one: autocatalysts (diesel-heavy, now substituting back into gasoline), jewellery (~a fifth of demand, China-led), industrial uses (glass fibre bushings, chemical and petroleum catalysts, medical) and investment (the only PGM with meaningful bar, coin and ETF demand). The hydrogen economy’s metal, if it arrives.
• Palladium — the specialist: ~80%+ of demand is gasoline autocatalysts, which made it the pure play on the combustion engine — a blessing through the 2016–22 supercycle, a curse since.
• Rhodium — the tiny, violent one: ~85% autocatalyst (NOx control), a market so small (~30 t/yr) that modest imbalances produced the wildest price ride in commodities ($700 → ~$29,800 → ~$4,700/oz across 2016–2024). No futures, dealer-quoted only.
• The minors — iridium (electrodes, PEM electrolysers — the hydrogen bottleneck metal), ruthenium (hard disks, chemical catalysts) and osmium (negligible market).

Geology concentrates the family absurdly: South Africa’s Bushveld Complex — a 2-billion-year-old layered intrusion — hosts ~83% of world reserves in two thin, deep horizons (the Merensky Reef and the chrome-rich UG2), mined kilometres underground by hand-held drills in the hardest mining conditions in the industry. The only other primary source of scale is Zimbabwe’s Great Dyke; Russia’s Norilsk produces its (palladium-rich) PGMs as a by-product of nickel-copper mining — supply that arrives regardless of PGM prices — and North America adds small primary (Montana’s Stillwater, currently curtailed) and by-product output. The value chain is long and slow: ore → concentrator → smelter → the converter plant and base/precious metal refineries — a pipeline holding months of metal, whose outages (the Anglo converter plant failures of 2020) can move world prices overnight.

Figure 2. The PGM value chain — two sources, one basket

Source: industry value-chain primers; conceptual diagram. The basket then flows to autocatalyst, jewellery, industrial, investment and hydrogen end uses (Section 2.2).

1.2 Units & measurement conventions

PGMs trade in troy ounces (oz) like all precious metals — 1 kg = 32.15 oz — while agency statistics (USGS) use kilograms and tonnes of contained metal; this report quotes prices per troy ounce and volumes in tonnes (world mined platinum ~170 t ≈ 5.5 Moz). The industry’s defining convention is the basket: South African miners report production and grades as 4E (Pt+Pd+Rh+Au) or 6E (adding Ir+Ru) ounces, and their revenue is the basket price — the prill-split-weighted average of the individual metal prices. The prill split (each mine’s fixed Pt:Pd:Rh ratio — Merensky ~60% Pt-rich, UG2 more Rh, Norilsk ~3:1 Pd:Pt) is the PGM analogue of a rare-earth deposit’s basket composition: it decides which price cycle a producer actually rides. Purity conventions: refined metal trades as sponge (industrial powder) or ingot/plate (99.95%+); rhodium and the minors trade dealer-to-dealer with no exchange at all. Flow vs stock: mine production and recycling are flows; above-ground stocks (AGS) — vaulted metal, ETF holdings, Chinese imports sitting in inventory — are the level that buffers deficits, and its measurement (WPIC publishes months-of-demand-cover estimates) is central to the investment case.

Table 1. PGM units and conventions

Source: USGS PGM Statistics , WPIC and LPPM conventions, 2025–26.

Numbers intuition: one catalytic converter holds 2–7 grams of PGMs (hybrids more, big diesels most); a tonne of Bushveld ore yields ~4–6 grams of 4E metal — meaning a car’s converter concentrates roughly a tonne of mined rock; the whole platinum market (~8 Moz demand) is worth only ~$10–12 billion a year at 2025 prices, and rhodium’s a tenth of that — small enough that one automaker’s inventory decision moves the price.

1.3 Pricing & benchmarks — fixes to dealer quotes

PGM pricing runs on a precious-metals rail with industrial-metal plumbing. Platinum and palladium have full market infrastructure: twice-daily LBMA/LPPM auction prices (the reference for physical contracts), liquid NYMEX/CME futures (the speculative and hedging venue), and OTC dealing in Zurich and London; sponge for industrial delivery trades at small premiums to ingot. Rhodium, iridium and ruthenium have none of it — they are dealer-quoted markets (Johnson Matthey and BASF base prices the industry references), thin, opaque and capable of moving 10% in a day, which is precisely why rhodium’s cycles reach five figures. Two structural features shape all PGM pricing. First, long-term contracts with automakers and industrials move most physical metal at index-linked prices, so spot swings pass through to producer revenue quickly. Second, the basket disciplines everything: South African supply responds to the combined 4E revenue, not to any single metal — rhodium’s 2020–21 spike alone kept whole loss-making platinum shafts alive, a cross-subsidy no single-metal analysis captures.

Table 2. Key PGM price benchmarks

Source: LPPM , CME contract specifications, Johnson Matthey price pages , 2025–26.

The long price story is really three stories. Platinum: a supercycle to $1,721 average in 2011 (spot $2,276 in March 2008), then a lost decade — dieselgate, substitution, ETF disillusion — grinding to $850–970 averages for eight straight years (2018–2024), before the deficit era ignited 2025: +84% from April to October, a $2,414/oz all-time record on 26 December 2025, and a ~$1,200 full-year average. Palladium: the mirror image — a $200 average in 2003 after the great Russian-supply panic of 2000–01, then the gasoline-engine supercycle from 2016 (average prices tripling to $2,419 by 2021, spot peaking at $3,440 in March 2022 on Russia fears) and the electric-era collapse to $995 by 2024, below most Western producers’ costs. Rhodium: $700 (2016) → ~$20,254 average in 2021 (spot ~$29,800) → $4,660 (2024) — the purest boom-bust chart in metals. The table carries the two majors; the minors’ landmarks live in the prose and Table 3’s source note.

Table 3. Average annual platinum & palladium prices, 2000–2025 (USD/oz)

Source: Engelhard/S&P Global Platts annual averages as compiled by the USGS Mineral Commodity Summaries — 2021 (2016–2020) and 2026 (2021–2025) editions; 2000–2015 from the Johnson Matthey/LPPM annual average record (selected years shown for space). Rhodium averages for context: ~$700 (2016), ~$20,254 (2021), ~$4,660 (2024), ~$5,800 (2025e). Spot landmarks: Pt $2,276 (Mar 2008), $2,414 record (26 Dec 2025); Pd $3,440 (Mar 2022).

Figure 3. Platinum & palladium prices, 2000–2025 (USD/oz)

Figure data: Table 3.

2. Supply, demand & the market balance

2.1 Where PGMs are mined — deposits & geology

PGM supply is geology’s most extreme concentration bet. The Bushveld Complex — a saucer of layered igneous rock the size of Ireland north of Johannesburg — holds ~83% of the world’s PGM reserves in two reefs rarely more than a metre thick: the platinum-rich Merensky Reef (the industry’s historic backbone) and the UG2 chromitite (now most of the ore, rhodium-richer but chrome-contaminated). Because the reefs are thin, deep (many shafts work 1–2 km down) and hard, most Bushveld mining remains labour-intensive, hand-drilled stoping — the highest-cost, most safety- and power-sensitive mining model in this series, chronically exposed to Eskom’s electricity failures and South Africa’s five-year wage cycles. The one great exception is Mogalakwena on the Bushveld’s northern limb — the world’s largest open-pit PGM mine and its lowest-cost major operation — with Ivanhoe’s Platreef ramping up nearby on the same thick-reef style. Zimbabwe’s Great Dyke (Zimplats, Unki, Mimosa) replicates Bushveld geology in friendlier mining widths; Norilsk in Arctic Siberia produces the world’s largest palladium stream from nickel-copper ores; and Montana’s Stillwater — the only US PGM mine, and the highest-grade major deposit anywhere — sat partly on care-and-maintenance through 2024–25, a casualty of the palladium crash.

Table 4. Mined PGM supply by country, 2025 (estimated)

Source: USGS Mineral Commodity Summaries 2026: Platinum-Group Metals , February 2026. Rounded; South African output fell ~9% and Russian ~6% in 2025.

Figure 4. Mined platinum & palladium by country, 2025 (t)

Figure data: Table 4.

Table 5. Major PGM mines & complexes (approximate recent annual output)

Source: company reports and production statements, 2024–25; output approximate on differing basket bases (4E/6E/2E) as noted.

2.2 Demand & consumption

PGM demand is the story of the catalytic converter, its heirs, and its hedges. Autocatalysts dominate: roughly 40% of platinum, over 80% of palladium and ~85% of rhodium go into exhaust systems, which ties the complex to global vehicle production (~90 million units), drivetrain chemistry (gasoline uses Pd-Rh; diesel Pt-heavy) and — decisively now — the electrification mix. Pure battery EVs use no PGMs, but hybrids use more than standard cars (cold-running engines need bigger catalysts), and the 2024–25 slowdown in Western BEV adoption in favour of hybrids materially extended the autocatalyst era — one reason palladium’s collapse found a floor and platinum’s deficit deepened. Around the auto engine: jewellery (~20% of platinum, a China-led market reviving on the gold-platinum price gap), industrial demand (glass-fibre bushings, paraffin/chemical catalysts, nitric-acid gauze, electronics — ruthenium’s hard disks, iridium’s electrodes), investment (platinum bars, coins and ETFs — the demand leg that exploded in 2025’s rally), and the long-dated wildcard, hydrogen: PEM electrolysers and fuel cells are Pt-and-Ir technologies, a potential million-ounce demand pillar in the 2030s that the market alternately prices in and forgets (2025’s iridium softness reflected exactly that fatigue).

Table 6. Platinum demand by segment vs palladium, 2024–25 (approximate shares)

Source: WPIC Platinum Quarterly segment data and Johnson Matthey PGM Market Report conventions, 2024–25; shares approximate, rounded.

Figure 5. What PGMs are used for — platinum vs palladium, 2024–25

Figure data: Table 6.

Geographically, PGM demand follows the world’s car plants and refineries: China is the largest consumer (autocatalysts, glass, jewellery), with Europe (diesel legacy), North America and Japan the other pillars — but consumption statistics matter less here than in base metals, because fabricated PGMs travel inside vehicles and equipment. The demand geography that does matter is China’s physical pull: since 2023 Chinese imports of platinum have persistently exceeded identified domestic demand, absorbing Western stocks into opaque inventories — one of the quiet forces behind the 2025 squeeze.

2.3 Supply: attrition, by-products & recycling

Primary PGM supply is in structural decline, and the reasons differ by source. South Africa (—9% in 2025) suffers the industry’s deepest cost disease: real electricity prices up multiples over a decade, Eskom load-shedding, ageing deep shafts, and a workforce-heavy model — through the 2023–24 basket-price slump roughly half of SA production ran at a loss, triggering shaft closures, deferred capex and consolidation warnings of “terminal decline” from the industry’s own lobby. Russia (—6% in 2025) mines palladium as a nickel by-product — supply that ignores PGM prices but inherits nickel’s troubles (grades, equipment access under sanctions, smelter maintenance). North America illustrates the marginal economics: Sibanye’s Stillwater — the West’s flagship — went partly onto care-and-maintenance in September 2024 despite US tax-credit support, and Impala’s Canadian Lac des Iles faces early closure. The stabiliser is recycling: ~140 t of Pt+Pd recovered globally in 2025, overwhelmingly from scrapped autocatalysts — a price-elastic, above-ground “mine” that sagged with low prices and cheap scrapped-car flows in 2022–24 and recovers with them. Add it up: total platinum supply has undershot demand three years running, and mined supply cannot respond quickly at any plausible price — deep-level shafts take a decade to build and a year to restart.

Table 7. World PGM supply structure, 2025e

Source: USGS MCS 2026 (mine production; global recycling ~140 t Pd+Pt in 2025) and WPIC supply detail; approximate, rounded.

Table 8. PGM reserves by country, 2025 (t of PGM content)

Source: USGS MCS 2026 , February 2026.

Figure 6. Where PGM supply actually comes from, 2025

Source: derived from Table 7 (USGS / WPIC ); approximate.

2.4 The supply–demand balance

Uniquely among this blog’s basket commodities, PGMs have professional balance sheets: the WPIC’s Platinum Quarterly and Johnson Matthey’s market reports publish audited-quality supply-demand series — and they currently tell the sharpest deficit story in metals. Platinum swung from pandemic-era surpluses to a run of deep deficits: roughly −0.9 Moz in 2023, −1.0 Moz in 2024, −1.1 Moz in 2025, and a forecast fourth consecutive deficit near −0.3 Moz in 2026 — cumulative shortfalls above 3 Moz in a ~8 Moz market, bridged entirely by drawing down above-ground stocks, which the WPIC projects will fall to under three months of demand cover by end-2026 (from over a year’s worth in 2021). That stock depletion — not a demand boom — is the mechanical engine of the 2025 repricing. Palladium ran contrary to its obituary: supply cuts (Stillwater, SA attrition, Russian slippage) outpaced the EV demand erosion, leaving 2024 in a ~0.5 Moz deficit (10.1 Moz demand vs 9.6 Moz supply) and the market near balance into 2026 — surpluses return only as electrification resumes eating autocatalyst demand. Rhodium remains structurally tight (SA-dependent, no elasticity anywhere). The buffer maths defines the regime: when AGS cover is measured in years, deficits are academic; measured in weeks, every quarter’s shortfall is a price event.

Table 9. The platinum balance & the stock drawdown, 2021–2026f

Source: WPIC Platinum Quarterly series and 2026 forecast releases (2025 deficit 1,082 koz; 2026f deepened to ~297 koz; AGS to just under three months by end-2026). Earlier years approximate and subject to WPIC revision — the trajectory, not the decimal, is the point.

Figure 7. Platinum’s deficit era, 2021–2026f

Figure data: Table 9. Above-ground stock cover falls alongside these deficits — from ~13 months of demand in 2021 to under 3 by end-2026.

Table 10. The three metals’ balances at a glance, 2024–26

Source: WPIC ; palladium 2024 per market reporting (supply ~9.6 Moz vs demand ~10.1 Moz); qualitative synthesis.

2.5 Trade flows & the refining chokepoint

PGMs move by armoured air freight, not bulk carrier, so their trade map is about processing capacity and counterparty risk, not shipping lanes. The physical flows: South African and Zimbabwean concentrate and matte through the SA smelting-refining complex (the world’s PGM chokepoint — when Anglo’s converter plant failed twice in 2020, palladium and rhodium repriced globally within days); refined metal to Zurich and London vaults, to Japanese and Chinese fabricators, and to the US; Russian palladium historically to Zurich and the West, since 2022 re-routed heavily toward China (unsanctioned, but self-sanctioned by many Western buyers and hit by the 2024 UK/US exchange bans on new Russian metal); and a reverse flow of scrapped autocatalysts from the US and Europe (the great car-scrapping basins) to refiners in Belgium, Germany, Japan and the US. China’s role deserves its own line: the largest fabricator, an aggressive stock-builder since 2023, and — through Zimbabwean and prospective Russian ties — an increasingly integrated buyer of the complex. The chokepoints, in order: SA smelters/refineries (capacity + Eskom), Russian export routes (sanctions architecture), and the recycling collection chain (dispersed, price-sensitive).

Table 11. Major PGM trade roles

Source: USGS import data, WPIC and trade reporting, 2024–26.

2.6 Market organisations & supply coordination

PGMs have no cartel today, but they are one of the few metals with a genuine cartel history: through the 1970s–80s the big South African houses (Rustenburg, Impala under the old Anglo/Gencor system) managed platinum sales through producer prices and allocation — a discipline that died with democratised markets and Russian supply. The modern coordination map: corporate concentration (three companies — Valterra, Impala, Sibanye — control most SA supply, and their shaft-closure decisions are the de facto supply policy); Nornickel as the single palladium superpower (~40% of mine supply); the South African state as an unintentional coordinator (Eskom’s power rationing has cut more supply than any OPEC meeting); and Russia’s export politics. The institutional layer: the LPPM (London market standards and the daily prices), the WPIC (producer-funded investment-demand development — the body behind the ETF ecosystem and the balance data), Johnson Matthey’s market reports, and the IPA (industry association). Government stockpiles are trivial (the US DLA holds kilograms, being disposed) — the strategic buffer is private AGS, which is exactly why its depletion matters.

Table 12. Who shapes the PGM market

Source: LPPM , WPIC , company disclosures, 2025–26.

3. The companies & the value chain

3.1 The largest PGM companies

The producer list is short, South African-heavy, and freshly reshuffled. Valterra Platinum — Anglo American Platinum until its demerger from Anglo completed in June 2025, with a Johannesburg primary and London secondary listing — is the value leader: Mogalakwena’s open-pit economics plus Amandelbult and the Mototolo/Der Brochen complex, ~3.5 Moz of PGMs a year, spun out (with impeccable timing) weeks before platinum’s rally ignited. Impala Platinum is the volume leader (~3.4 Moz 6E across Rustenburg, Zimplats, Two Rivers, Bafokeng and Canada — the last now facing closure), the classic deep-level operator. Sibanye-Stillwater spans SA PGMs (Marikana, Rustenburg, Kroondal), the curtailed US Stillwater mines and gold — the sector’s highest-beta major, whose stock round-tripped the entire palladium cycle. Northam Platinum is the growth-focused independent (Zondereind, Booysendal); Nornickel towers over palladium as a nickel company; Ivanhoe’s Platreef is the next major entrant; and Tharisa (chrome-PGM co-product) plus the Zimbabwean operators fill the list. Note the ownership churn as a signal: Anglo demerged, Sibanye assembled distressed assets a decade ago, Implats absorbed Bafokeng — PGM assets change hands at the cycle’s extremes, and the last two years were an extreme. Companies are compared on production, resources and cost — never market capitalisation.

Table 13. Leading PGM companies, 2024–25 (approximate)

Source: company production reports 2024–25 (Valterra demerger ; Impala, Sibanye, Northam, Nornickel disclosures); output approximate on differing basket bases (PGM/6E/4E/2E). Listing notes: Nornickel trades in Moscow under sanctions constraints; Zimplats is Impala-controlled. No market-capitalisation figures are shown by design.

3.2 Company archetypes along the value chain

PGM business models sort by geology, basket and stance toward the cycle. The open-pit/thick-reef producer (Mogalakwena, Platreef, Great Dyke operations) owns the structural low ground: mechanised, safer, expandable — the archetype the industry’s future concentrates into. The deep-level conventional miner (most of Rustenburg-belt production) is the sector’s defining bet: enormous operating leverage to the basket price, chronic exposure to power, labour and depth — printing money at 2021 basket prices, closing shafts at 2024’s. The by-product producer (Nornickel, Sudbury) supplies regardless of PGM prices — the demand-side’s blessing and the price-side’s curse. Recyclers and refiners (Johnson Matthey, BASF, Umicore, Heraeus) earn spreads and capacity rents on the scrap loop. Chrome/PGM co-producers (Tharisa) hedge the basket with chromite. And the royalty/streaming layer exists here (unlike most of this series): Wheaton and Franco hold palladium-platinum streams on Stillwater and Sudbury — small, but real. The investable spectrum runs from Valterra (quality, low cost) through Impala (volume torque) to Sibanye (maximum beta, both directions — proven both ways in 2022–25).

Table 14. PGM company archetypes

Source: company filings; the Metal Pilot project-type taxonomy, 2025.

Figure 8. PGM company archetypes by price sensitivity

Source: conceptual, see Table 14.

3.3 Infrastructure & balance-sheet assets

Reading a PGM producer’s filings means valuing a pipeline as much as a mine. The resource is described by reef type (Merensky/UG2/Platreef), 4E grade (g/t) and — decisively — the prill split, which fixes the revenue basket; depth and mining method translate directly into cost structure and safety/power exposure. Downstream, the processing chain is the moat: concentrators, smelters (furnace availability is a production constraint — matte throughput, rebuild schedules), the converter plant (the ACP’s 2020 double failure taught the market its importance) and base/precious-metal refineries — with months of pipeline inventory inside them, so a producer’s sales and production can diverge for quarters (release of pipeline metal flattered several 2024 results). PGM-specific balance-sheet lines: work-in-progress metal (a genuine store of value), Eskom dependence and self-generation build-out (solar farms are now core capex), five-year wage agreements (the labour cycle), deep-shaft closure liabilities, and — on the revenue side — the rand/dollar exchange rate, since costs are rand and revenue dollars: a weak rand has repeatedly rescued the industry’s margins (and vice versa).

Table 15. PGM company asset types and metrics

Source: company disclosures and technical reports (SAMREC/JORC), 2024–25.

4. Investing in PGMs

4.1 Ways to own them — exposure vehicles

PGMs are the first family in this series with a full precious-metals ownership stack. Physical metal is genuinely ownable: platinum (and palladium) bars and coins (American Eagles, Maple Leafs) trade like small gold products, with higher dealing spreads; rhodium exists physically only via niche bars with punishing spreads. Physically backed ETFs are the institutional route — platinum ETFs (WPIC-fostered, several Moz held globally) own vaulted metal with no roll mechanics, and their flows are themselves a market force (the 2025 rally drew record inflows). Futures (NYMEX PL/PA) serve leverage and hedging — with the standard futures curve caveats (roll and margin mechanics) applying to futures-based strategies, though physical ETFs sidestep them entirely. Equities provide the torque: SA majors (Valterra, Impala, Sibanye ADRs) are effectively leveraged baskets of the PGM prices divided by the rand, with 2–4× the metal’s moves in both directions; streamers (Wheaton’s Stillwater stream) offer capped-risk exposure. And uniquely: rhodium, iridium and ruthenium are essentially uninvestable directly — the only practical access to their cycles is through the SA miners’ baskets, which is worth knowing when rhodium runs.

Table 16. Ways to own PGMs

Source: issuer and exchange documentation (WPIC on platinum investment , CME specifications); structural attributes only — no fees or prices, which change.

4.2 How to value & screen PGM producers

PGM screening is basket arithmetic on South African stress. Start with the basket, not a metal: compute each producer’s realised 4E basket price from its prill split, then set it against AISC per 4E ounce — the sector’s margin line — and stress it at trough baskets (2024’s ~$1,200/4E oz sent half of SA mining underwater) and at current ones. Then the prill-split positioning: platinum-rich splits (Mogalakwena, Zimbabwe) ride the deficit story; palladium-rhodium-heavy splits (UG2 deep-levels, Stillwater) ride the autocatalyst sunset — same industry, opposite bets. Then the South African stress tests: power exposure (self-generation share vs Eskom), labour cycle timing (wage agreement expiries), shaft age and depth (the capex cliff), and smelter pipeline health. For all: rand sensitivity (a 10% ZAR move swings margins like a 10% basket move), balance-sheet endurance (survivors of 2024 earned the right to 2025), and the recycling/streaming alternatives for lower-risk expressions. These are the screens — production, basket quality, cost position, resources — you can run across every listed PGM producer on Metal Pilot.

Table 17. PGM producer screening metrics

Source: company filings, 2023–25; thresholds indicative from the 2023–24 trough experience.

Figure 9. Illustrative PGM cost curve (AISC per 4E oz vs cumulative supply)

Chart source: illustrative; cost bands from company disclosures, 2023–25 — approximate, not company-level data. The two dashed lines mark the 2024 trough basket ($1,200/4E oz) and the 2025 rally basket ($1,700/4E oz).

4.3 Macro regimes, rates & correlations

PGMs live a double life: precious metals in the portfolio, industrial metals in the physics. The precious side ties platinum to gold — real rates, the dollar, safe-haven flows (2025’s rally fed partly on rotation from record-priced gold into its cheap sister metal, and the gold-platinum ratio is a classic mean-reversion watch). The industrial side ties the complex to the global auto cycle — vehicle production, drivetrain mix, Chinese consumption — and to South African supply shocks, which repeatedly overwhelm macro logic (the 2008 power crisis spiked platinum into a global recession). Regime summary: PGMs do best in precious-metals bull phases combined with auto-cycle strength or supply failure, worst in auto recessions and EV-acceleration scares, and idiosyncratically in between. (For the cross-commodity framework, see the macro regime guide .) Rate sensitivity is real but weaker than gold’s — the industrial demand base anchors them — and the rand matters as much as the dollar for the equities.

Table 18. PGMs across regimes

Source: price record (Table 3) and episode history; regime characterisations are historical, not predictive.

On past performance, the complex offers the cleanest matched pair of boom and bust in modern commodities. The equity amplifier, concluded: palladium’s collapse from its March 2022 peak ($3,440 spot; $2,419 2021 average) to a $995 average in 2024 — roughly −60% on annual averages, −70%+ peak-to-trough — took Sibanye-Stillwater’s NYSE listing from ~$20 to under $4 (−80%+), forced the Stillwater curtailment and billions in impairments, and dragged Impala down comparably; the deep-level cost base turned a metal bear market into an existential one. The reversal then paid the survivors: platinum’s 2025 run (+84% April–October, +~25% on annual averages) roughly tripled the SA miners’ equities inside a year — Sibanye’s CEO professed himself surprised by his own stock. Longer-run, the lesson of 2000–2025 is that PGMs reward cycle entry, punish buy-and-hold in the wrong metal (palladium at the 2001 peak took two decades to revisit; platinum bought in 2008 waited seventeen years) — and that the basket miners embody both cycles at once. Past performance is not indicative of future results.

On correlations (monthly, 2010–2025): platinum runs +0.6 with gold (the precious anchor) and +0.5 with silver, moderately with the industrial complex; palladium correlates more with the auto cycle and less with gold (+0.3); both are negative on the dollar (−0.4); and rhodium correlates with nothing except South African supply and Chinese ordering. The equities add a + rand-gold-beta structure of their own. In supply-shock episodes, all cross-correlations collapse — the complex trades on its own physics.

Table 19. PGM correlations (monthly, 2010–2025, approximate)

Source: author analysis of LPPM/Platts price series against FRED macro series, monthly, 2010–2025. Approximate; correlations break down in supply-shock episodes.

Figure 10. Platinum correlations, monthly 2010–2025

Figure data: Table 19. South African supply and auto-drivetrain news is the episodic driver that overrides these correlations in supply-shock periods (Section 4.3).

4.4 Price drivers & cycles

The durable drivers, read through concluded episodes. Demand side: vehicle production and drivetrain mix (the master variable — gasoline/diesel/hybrid/BEV shares each reweight the three-metal demand basket), emissions regulation (every tightening, from Euro standards to China 6, has been a demand event), substitution and thrifting (automakers re-engineer loadings within 2–3 years whenever a metal runs — the force that capped every spike), and investment flows for platinum. Supply side: South African operational reality (power, labour, depth), Russian logistics and politics, the processing pipeline (smelter/ACP outages), and recycling flows. The settled case studies: the 2000–01 palladium crisis — Russian shipments halted, prices quintupled, and Ford, panic-buying at the top, wrote off ~$1 billion when prices collapsed — the founding lesson in inventory panic. The 2008 double shock — South Africa’s January power emergency spiked platinum to $2,276 before the financial crisis crushed it to ~$760 within months: supply shock, meet demand shock. Dieselgate (2015) — a regulatory scandal that structurally re-weighted demand from platinum (diesel) to palladium (gasoline), launching palladium’s five-year supercycle and platinum’s lost decade. The ACP failures and rhodium mania (2020–21) — two converter-plant explosions bottled up SA output into COVID-recovering demand; rhodium touched ~$29,800. And the palladium supercycle’s end (2022–24) — EV fear plus destocking unwound six years of gains in two. The 2025–26 platinum squeeze is the live chapter — deficits, stock depletion, the EU’s combustion-ban U-turn — flagged here as current context, not settled history.

Table 20. PGM price drivers

Source: WPIC , USGS and the episode record. Case studies are concluded episodes; the 2025–26 squeeze is flagged as unresolved.

4.5 Risks, controversies & ESG

The bear cases are specific. Terminal-demand risk is the big one: autocatalysts are a technology in managed decline, and a faster-than-expected BEV re-acceleration would hit palladium and rhodium first and hardest — the 2022–24 collapse was a preview, not the finale. Substitution risk cuts within the family (automakers switching Pt↔Pd caps relative rallies) and beyond it (fuel-cell and electrolyser designers thrifting iridium and platinum loadings). Concentration risk is double-edged: South Africa’s problems restrict supply (bullish) while threatening the industry’s viability (existential — the sector’s own lobby warns of terminal decline in deep-level mining). Russia remains a sanctions wildcard for palladium.

The ESG ledger is dominated by the human dimension of deep-level mining. Marikana (August 2012) — police killed 34 striking miners at Lonmin’s operations, now Sibanye’s — remains the industry’s darkest modern chapter and the reference point for its labour relations; safety in kilometre-deep, hand-drilled stopes remains among mining’s hardest problems, though fatality rates have fallen markedly. Community, migrant-labour legacy and inequality issues around the Rustenburg belt persist; power scarcity ties the industry to Eskom’s coal fleet (and pushes it, constructively, into large-scale solar). Set against this: PGMs are pollution-control metals by function — billions of tonnes of avoided vehicle emissions — recycling is world-class (a scrapped converter is among the most recycled objects on Earth), and the hydrogen economy, if it scales, makes them decarbonisation enablers. These are contested questions, and reasonable analysts weigh them differently.

Figure 11. PGM risk map — likelihood vs. impact

Source: author’s qualitative assessment; see Section 4.5.

5. Future outlook & forecasts

PGM forecasting is drivetrain forecasting, and the honest frame is a race between two decays: demand’s (autocatalysts fading with electrification) and supply’s (South African deep-level attrition). Forecasts are scenarios, not measured facts — and the 2020s have already humbled both bulls (palladium) and bears (platinum).

5.1 Demand

The consensus shape: autocatalyst demand declines gently, not suddenly — hybrids (PGM-heavier) take share from both pure ICE and BEV trajectories, emissions rules tighten loadings in India and China, and the EU’s 2025 U-turn on its 2035 combustion ban extended the runway. Platinum’s diversification helps it most: jewellery revives on its discount to gold, industrial demand compounds quietly, investment demand has become structural through ETFs — and hydrogen offers the only million-ounce-scale new demand in the complex (PEM electrolysers and fuel cells), albeit repeatedly deferred. Palladium and rhodium remain leveraged, undiversified bets on the combustion engine’s twilight pace. The WPIC projects platinum demand near 8 Moz through the decade with deficits persisting into 2026 (~−0.3 Moz, the fourth consecutive).

5.2 Supply and the balance

Supply’s trajectory looks locked: South African output declines (no new deep-level shafts are being built; closures outpace Platreef-class additions), Russia drifts lower, recycling recovers cyclically but cannot re-mine what thrifting removed, and above-ground stocks — the historic buffer — approach functional depletion for platinum (under three months’ cover by end-2026). That is the structural core of the bull case: even flat demand meets falling supply. The bear case needs a demand cliff (BEV re-acceleration) or a stock surprise (opaque Chinese and Russian inventories re-emerging). Palladium balances near-term on supply cuts, then tips to surplus as electrification resumes — with its own floor set by how much SA and US supply exits first. The wildcards: hydrogen’s pace, Russian metal’s routing, and whether 2025’s price levels finally fund the mechanised-mining renewal South Africa needs.

Table 21. PGM outlook indicators (scenario figures)

Source: WPIC Platinum Quarterly and 2026 forecasts ; USGS MCS 2026 . Scenario projections, not measured data.

5.3 Catalysts to watch

Near term: WPIC quarterly balances and AGS estimates (the scoreboard), NYMEX/ETF flows and lease rates (the squeeze gauges — lease-rate spikes flagged every leg of 2025), South African operational news (Eskom, wage rounds beginning, shaft decisions), drivetrain data (hybrid vs BEV share monthly), and Chinese import behaviour (the opaque accumulator). Corporate: Valterra’s first standalone years, Sibanye’s Stillwater restart decision (the marginal Western ounce), Platreef’s ramp, and consolidation among juniors. Structural: hydrogen electrolyser FIDs (the iridium-platinum option), emissions-standard rollouts in India/China, Russian sanctions architecture, and the gold-platinum ratio for rotation flows. What would confirm the bull thesis: continued deficits into genuinely depleted stocks with SA supply still falling. What would break it: a BEV re-acceleration, a Chinese stock disgorgement, or basket prices high enough for long enough to reopen shafts — the cycle’s own cure.

Table 22. PGM catalyst calendar

Source: WPIC , company calendars, market data providers.

6. Summary

The platinum group metals are six rare sisters mined at scale in just two geological freaks — South Africa’s Bushveld (~83% of reserves, ~70% of mined platinum) and Russia’s Norilsk (~40% of palladium, as a nickel by-product) — plus a recycling loop worth a quarter of supply, and consumed mostly by the catalytic converter (~40% of platinum, ~80%+ of palladium and rhodium). Pricing runs precious-style — LPPM fixes and NYMEX futures for the majors, dealer quotes for violent little rhodium — and the price record is the wildest in the precious complex: palladium’s $200 (2003) → $3,440 (2022) → $995 (2024) round trip; rhodium’s $700 → $29,800 → $4,700; and platinum’s lost decade at $850–970 ending in 2025’s ~84% rally to a $2,414 record, powered by the WPIC’s deficit era — four consecutive shortfalls (−0.9, −1.0, −1.1, ~−0.3 Moz) draining above-ground stocks toward under three months of cover. Supply is in structural attrition: South Africa’s deep, power-starved shafts closed through the 2023–24 trough (the sector’s own lobby warns of terminal decline), Stillwater sits curtailed, and no new deep-level capacity is coming. Demand declines more slowly than feared: hybrids extend the autocatalyst era, platinum substitutes into palladium’s seat, jewellery and industry hold, and hydrogen waits in the wings. The companies — Valterra (quality, freshly demerged), Impala (volume), Sibanye (maximum beta, −80% then 3× inside four years), Nornickel and the Zimbabweans — are screened on basket price vs AISC per 4E ounce, prill split, power self-sufficiency and rand exposure, never market cap. The regime profile is a precious-industrial hybrid that trades its own supply-shock physics. The single most important variable to watch is the pace of South African supply attrition against the drivetrain transition, with above-ground stock depletion as the fuse.

To go from this big-picture view to the actual companies — screening every PGM producer by production, resources, basket and cost — explore Metal Pilot.

7. Vocabulary

Every specialised term used in this report, defined once at first use in the body, collected here for reference.

Table 23. PGM vocabulary

Source: definitions follow LPPM , WPIC and industry terminology, 2025–26.

8. Sources, methodology & disclaimer

8.1 Sources, methodology & data vintage

Agencies & official data: USGS Mineral Commodity Summaries 2026: Platinum-Group Metals and the 2021 edition (price and production history); USGS PGM Statistics and Information ; FRED macro series.

Industry & market bodies: WPIC Platinum Quarterly and deficit-forecast releases (2025: 1,082 koz; 2026f , deepened to ~297 koz ); LPPM ; Johnson Matthey PGM market data; CME contract specifications; 2025 rally coverage (Discovery Alert/market reporting ).

Company filings & statements: Valterra Platinum annual results and demerger documentation ; Impala Platinum, Sibanye-Stillwater (incl. the Stillwater curtailment and Lac des Iles closure reporting ), Northam, Nornickel and Ivanhoe disclosures, 2024–25.

Methodology: prices are annual averages (Engelhard/S&P Global Platts as compiled by the USGS for 2016–2025; the Johnson Matthey/LPPM annual record for 2000–2015), never spot snapshots; spot landmarks are dated explicitly. Mine production and reserves follow the USGS; the platinum balance and above-ground-stock estimates follow the WPIC (earlier years approximate and subject to revision); palladium balance figures follow market reporting as cited. Company volumes mix basket bases (PGM/6E/4E/2E) and reporting periods as noted under each table. Correlations use monthly data 2010–2025 and are approximate. Reserves, resources and forecasts are estimates, not measured facts.

Data as of: 6 July 2026. Intended update cadence: annually after the USGS Mineral Commodity Summaries (February), with the WPIC’s quarterly balances as interim anchors.

8.2 Disclaimer & disclosure

This report is for informational purposes only and is not investment advice, a recommendation, or an offer to buy or sell any security or commodity. PGM prices are exceptionally volatile — the recent record includes a 70% collapse and an 84% rally within four years — and the figures here are estimates as of the stated date that will change; balances, stock estimates, correlations and regime observations are estimates and historical patterns that may not persist. Do your own research and consult a licensed financial adviser before acting. This report was prepared with the assistance of AI; its figures were sourced from the references above and reviewed, but readers should verify any number before relying on it. The author holds no position disclosed as a conflict in respect of the companies named.