Gold & Precious Metals Mining

Overview / thesis

Gold is in a structural bull market, and the producers are sitting on the highest dollar-margin per ounce in history. That is the whole thesis in one line. In April 2026 gold trades around $4,700/oz. Goldman Sachs has it at $5,400 by year end; Wells Fargo says $6,100–6,300; UBS sees $6,200 with upside to $7,200. These are not crypto-bro forecasts — they are sober institutional desks revising up after two years of being wrong on the upside. The base case from major bank desks puts gold $5,000–6,500 by end of 2026, another 10–40% from current levels.

The setup is unusual because three structural forces line up at once: demand is structurally rising (de-dollarization, central bank accumulation), supply growth is structurally capped (peak-gold geology), and producers carry record cash margins. World Gold Council math implies an industry margin of roughly $2,800 per ounce in 2026 against ~$1,500 AISC. That is the spread that turns a sleepy commodity sector into a five-year compounding story. The work across these sources is figuring out who actually captures it.

Why gold now — the three things that changed at once

1. Central banks are buying at structural, not cyclical, rates. Central banks bought 1,237 tonnes in 2025, the third consecutive year above 1,000 tonnes, against a pre-2022 average of 400–500 tonnes. The 2026 forecast is 755 tonnes — even at the lower number, that is double the pre-2022 norm. 76% of central banks plan to increase gold holdings over the next five years. China's PBoC has bought every month for 15+ consecutive months as of early 2026. Drivers: the Russia sanctions experience (reserves can be frozen), dollar dilution, and BRICS reserve diversification. This is the demand driver that recurs across every source as the load-bearing one.

2. The dollar is losing reserve share. The dollar's share of global reserves fell from 71% in 1999 to 56.3% in mid-2025, a thirty-year low. Gold is the historical hedge for both a weakening dollar and falling real rates, and both are moving in gold's favor.

3. Peak-gold geology caps supply. Mine production hit a record 3,672 tonnes in 2025, yet barely keeps pace with demand. No major deposit (>2 million ounces) has been discovered since 2022. Reserves at the top 20 producers fell 26% from 2012–17 and have not fully recovered. New mine builds take 10–15 years from discovery to first pour. The Moore's Law equivalent for gold mining runs in reverse: average reserve grade for major producers has fallen from ~1.5 g/t in 2000 to ~1.0 g/t today. Supply growth has run 1–2% per year for a decade and is forecast to peak around 2027–2028, then plateau or decline as the discovery shortfall bites. Anyone holding permitted ore in a Tier-1 jurisdiction at ~$1,500 AISC owns a moat that is essentially uncopyable for a decade.

Sizing the opportunity (TAM and revenue pool)

Global gold mine production was 3,672 tonnes in 2025, worth ~$390 billion at the $3,300 average. At $4,700 spot the producer-set run-rate revenue is closer to $550 billion. Production growth has averaged 1–2% per year and is structurally capped through 2030. The profit pools by layer of the value chain (2026E):

Layer	Approx revenue pool (2026E)	Gross margin	Concentration
Exploration (juniors)	$5–8B	Negative	Highly fragmented
Mining equipment OEMs	$90–100B	30–40%	Oligopoly (Cat/Komatsu/Sandvik/Epiroc)
Engineering & construction (EPCM)	$20–30B	5–10%	Fragmented
Producers (the miners)	$200B+	40–60% at current gold price	Top 10 = 30% of supply
Refiners (LBMA)	$1–2B in fees	10–20%	Switzerland-dominated
Bullion banks (market makers)	$5–10B	High	Top 5 dominate
Royalty & streaming	$4–6B	70–80%	Top 3 = 80%
Vaulting & custody	$1–2B	40–50%	Concentrated

The most attractive economics, layer by layer, are the royalty/streaming companies (highest margin, lowest operational risk), the producers (the leveraged play on gold price), and the equipment oligopoly (steady, secular). Least attractive: exploration juniors (lottery tickets) and EPCM contractors (commoditized). End-use demand splits roughly: central banks ~25%, jewelry ~50% (India + China dominant), ETFs and bars/coins ~20% (the swing factor), industrial ~5%.

The 'so what' — record margins meet a discipline test

The producer set is fragmented (top 10 = ~30% of global supply; the largest single producer, Newmont, has ~6% share — there is no Aramco of gold), so producers have zero pricing power on gold itself. Their only levers are cost control and capital allocation. That makes the bull market a test of management behavior, not just price: with cash margins at record highs, will teams return capital, or buy each other at the top of the cycle again?

The 2011–2015 bear market killed the M&A-junkie super-majors — Barrick, Newmont, Goldcorp all wrote down billions after paying up for assets at $1,800 gold that then fell to $1,050 in 2015. The post-2016 era was about discipline: cut capex, focus on margin, return cash. The companies that learned the lesson (Agnico the obvious example) are now the quality compounders; the ones that didn't still trade at mid-single-digit P/E because investors don't trust them to spend the windfall. So the sector edge is to buy the disciplined operators in Tier-1 jurisdictions, not the index.

The producer-vs-spot gap (the tactical case)

Historically, a 33% move in gold ($1,800→$2,400) drove the GDX miners 60–80%. But the 2024–25 move from $2,000 to $4,700 (+135%) was matched by miners moving only ~100–120% — miners lagged spot for the first time in two decades. The drag: cost inflation eating margins early in the cycle, ESG capital flight, and crypto absorbing speculative flows. As the market internalizes that AISC has stabilized around ~$1,500 and gold is holding above $4,000, that gap should close. This is the explicit 12–18-month catch-up setup — equity leverage that hasn't been priced yet.

Cyclicality — both clocks point bullish

Gold mining runs two cycles at different timescales: the gold price cycle (5–7 years between major peaks) and the capital cycle (a 7–10 year lag between price highs and new mine completions). Right now both are aligned bullishly — high price, no new supply coming online for several years — which is why operating margins sit at record highs. The honest read on positioning: mid-to-late innings of a multi-year bull market that began in late 2022.

The key debate

Where the bull case can break, and what the sources argue:

• Does gold have to keep ripping? No — and this reframes the whole sector. The highest-conviction single name's brownfield-expansion economics work at $3,200 gold (base case), down to ~$2,500, not $4,500. The best producer theses are buying margin and growth, not a leveraged bet on continued price appreciation. But the overall stack of equity theses still depends on gold holding above ~$3,500; below that, the premium-valued names (and the royalty names priced for continuation) compress fastest.
• Real rates. A surprise tightening cycle is the single biggest macro risk — the 2013 taper tantrum is the playbook. Watch the 10-year TIPS yield: below 1%, gold rallies; above 2.5%, gold struggles.
• Cost inflation. Industry AISC rose from ~$1,000 in 2020 to ~$1,521 in 2025. If it keeps climbing, cash margins compress even at higher gold prices.
• Recycling elasticity. 2025 saw record recycled supply (1,330 tonnes). At $5,000+ scrap could become a meaningful headwind to net-new demand.
• Crypto cannibalization and ESG flight. Bitcoin has absorbed some "store of value" demand (caps retail upside but hasn't dented central bank or jewelry demand); European institutional pools excluding mining outright caps the producer multiple ceiling.
• Discipline vs. top-of-cycle M&A. The late-cycle sell signal is management teams announcing new greenfield builds or a wave of large super-major combinations. Small deals are healthy; mega-mergers are the warning.

Jurisdiction as the live screen — the "no-Africa" filter

The defining sector overlay in this body of work is jurisdictional risk, which used to discount valuations by 10–20% and in 2026 discounts them by 30–50%. West African gold mines have been seized by junta governments in Burkina Faso, Mali, and Niger over 2022–2025. The Fraser Institute Tier-1 list — Canada, Australia (WA), Nevada, Finland, Sweden — is where the durable assets sit. The screen that organizes the company-level work ("Canadian-listed, no Africa exposure") yields exactly five producers: AEM, AGI, EGO, WDO, LUG. The leading royalty complement is FNV. Company-level theses, sizing, and catalysts live on those ticker pages; the cross-cutting takeaway here is that in a structural bull market the alpha is in jurisdiction-clean, low-cost, capital-disciplined names, not broad sector beta.

Leading indicators to watch

1. Gold ETF holdings (GLD, IAU) — North American ETF buying turning positive after years of outflows is the signal institutional money is rotating in.
2. Producer capex announcements — when management starts saying "we're going to build new mines," that's the late-cycle sell signal.
3. Mining M&A activity — small deals healthy; a wave of large super-major combinations is the warning.
4. Real interest rates — the 10-year TIPS yield.
5. Central bank buying continuity — if China or Russia stops buying for two consecutive months, that's a major tell.

The one-paragraph thesis: gold is in a structural bull market driven by central bank buying, dollar reserve diversification, and supply-constrained geology, while producers carry record cash margins (~$2,800/oz at spot vs ~$1,500 AISC). The best way to play it is the producers and royalty companies that combine Tier-1 jurisdictions, low cost, and disciplined capital allocation — and miners have lagged spot, leaving equity leverage that hasn't yet been priced.

How it works

Gold exists as an industry because humans need a money asset that no government can print and no counterparty can default on. Every other store of value in history either inflated away (paper currencies), defaulted (sovereign bonds), or proved breakable (the 1971 dollar peg). Gold's unusual property is that the global stockpile grows by roughly 1.5% per year through mining — almost identical to long-run population growth — so per-capita supply stays nearly flat across centuries. It survives where salt, cowrie shells, and Mediterranean silver did not because two things are true at once: it is rare enough that even the huge 19th-century California, Australia, and Witwatersrand rushes only added single-digit percent to the existing stock per year, and it is chemically inert, so almost every ounce ever mined still exists somewhere. The demand for that property is permanent and the supply is hard in a specific way — hard to find, hard to extract economically, hard to refine to bullion-grade purity, hard to move and store. Each "hard" step is a separate profit pool.

The physics and chemistry of the metal

Gold is element 79. It is one of the densest metals known at 19.3 g/cm³ (almost twice lead), chemically nearly inert (it does not oxidize at room temperature and dissolves only in aqua regia or cyanide solutions), and rare in the crust at about 4 parts per billion on average. The average rock contains roughly 0.000004 grams of gold per ton. An economic ore body today runs anywhere from 0.5 g/t (huge low-grade open pit like Newmont's Boddington) to 20+ g/t (deep underground narrow vein like Wesdome's Eagle River). That spread represents up to a five-million-fold concentration over background crust, and it is the central fact that dictates every downstream engineering choice.

Two main geological processes do the concentrating, with a third derivative category:

• Hydrothermal deposits. Hot, mineral-rich water circulates through rock fractures, dissolves trace gold, and re-precipitates it where temperature, pressure, or chemistry change. This builds the epithermal and orogenic vein systems — narrow but high-grade ribbons of quartz and gold in older rock. Most underground mines target these: Wesdome Eagle River, the entire Abitibi greenstone belt in Quebec and Ontario, and Carlin in Nevada are all hydrothermal.
• Placer deposits. When hydrothermal veins erode over millions of years, gold particles (denser than surrounding sediment) wash downstream and concentrate in river gravels. This is what 1849 California, 1896 Klondike, and modern artisanal miners chase. Placer is only a few percent of global supply today but built the early industry.
• Paleoplacer. Essentially fossilized placer — ancient river systems buried under younger rock. South Africa's Witwatersrand basin is the world's largest, producing roughly half of all gold ever mined. It is also why the South African industry is dying: the easy material is gone, remaining grade is low, and depths over 4 km make it the most expensive mining on Earth.

The vocabulary that controls the economics

Five terms govern whether any deposit or quarterly result actually works:

• Grade (g/t, grams per tonne): how much gold per tonne of ore. Below 1 g/t is low grade (open pit, bulk tonnage); above 5 g/t is high grade (underground, narrow vein). Grade determines which mining method is economic.
• Recovery (%): of the gold present, how much ends up in the final doré bar. Modern flowsheets recover 85–95%. The remaining 5–15% is locked in sulfides or refractory minerals and is either lost or requires extra processing.
• AISC (all-in sustaining cost, $/oz): the all-in cash cost to produce one ounce — direct mining, processing, royalties, sustaining capex, and G&A, but excluding growth capex. Industry average in 2025 was $1,521/oz. Below $1,200 is excellent in 2026; $1,200–$1,500 is acceptable; above $1,800 means the company only works at high gold prices.
• Reserves vs Resources: reserves are gold proven economic at a specific gold-price assumption and are bankable; resources are inferred-but-unproven optionality. Reserve calculations carry a hidden lever — the gold price assumed when they were booked. AEM calculates reserves at $1,600/oz (35% of spot), which means its M&I + Inferred ounces (88.9 Moz combined) are huge upside optionality at spot prices.
• Doré: the impure gold-silver alloy bar (typically 70–90% gold, often cited as 80–90%) poured at the mine site. This is what ships to a refiner for purification to bullion grade.

The single most important derived number is AISC margin — gold price minus AISC, i.e. the cash margin per ounce. At $4,700 gold and $1,500 AISC, that is roughly $3,200/oz of gross cash margin. World Gold Council math implies an industry margin of roughly $2,800 per ounce in 2026 against $1,500 in AISC, the highest dollar-margin per ounce in history.

How a tonne of rock becomes a Good Delivery bar — the six stages

The full flow from rock-in-the-ground to delivered LBMA vault bar is six distinct stages, each with its own economics, technology, and suppliers.

Stage 1 — Exploration (5 to 15 years, junior miners). Geologists identify targets through remote sensing, geochemical sampling (panning streams, soil grids), and geophysics (magnetic, gravity, induced-polarization surveys), then drill thousands of meters of diamond core to map the ore body in 3D. A "discovery" usually means a few hundred thousand ounces inferred; a "deposit worth building" means a few million ounces with grade and metallurgy that pencil at conservative prices. Roughly 1 in 1,000 explored targets becomes a producing mine.

Stage 2 — Development (3 to 7 years, $500M to $2B+ capex). Once delineated and permitted (the slow part — environmental review, indigenous consultation, water rights, tailings approvals can take a decade), construction begins: stripping overburden in open pits, sinking shafts and developing levels underground, building the mill, tailings dam, camp, and power line. Capital intensity has worsened — a new million-ounce-per-year project today costs $1.5–2.5 billion. This is why seniors buy permitted juniors rather than build greenfield.

Stage 3 — Mining (the operating mine). Two main approaches:

• Open pit: bench-by-bench excavation of shallow, low-grade, large-tonnage deposits using massive haul trucks (Caterpillar 797F at 400-tonne capacity), electric rope shovels, and bulldozers. Strip ratio (waste tonnes moved per ore tonne) drives the economics. This is where Caterpillar, Komatsu, Hitachi, and Liebherr sell most of their gear.
• Underground: shafts, declines (spiral ramps for trucks), levels, and stopes; trackless equipment (load-haul-dump machines, jumbo drill rigs, articulated trucks) moves ore to a hoist or up a decline. Sandvik, Epiroc, and Cat hold roughly 75% of underground trucks and ~88% of LHDs.

Stage 4 — Milling and processing (the plant). This is where the chemistry happens and the heaviest energy is consumed. Ore is crushed (jaw crusher, then cone crusher), ground (semi-autogenous and ball mills — the most energy-intensive step in the whole flowsheet), then leached. The dominant leach for free-milling ore is CIL (carbon-in-leach):

Crushed ore → Ball mill → Cyanide leach tank (with carbon)
                              ↓
                    Loaded carbon (carries gold)
                              ↓
                    Strip vessel (high T, high P)
                              ↓
                    Gold-bearing electrolyte
                              ↓
                    Electrowinning cells
                              ↓
                    Gold sludge → Smelt → Doré bar

Cyanide added to the slurry dissolves gold into solution; activated carbon in the same tank captures the dissolved gold like a sponge; the loaded carbon is washed with hot caustic to release gold into a concentrated electrolyte; electric current at the electrowinning cell deposits gold as a metallic sludge on cathodes; the sludge is melted in an induction furnace and out comes the doré bar.

Three flowsheet variants change the cost structure entirely:

• Heap leach — for lower-grade or oxide ores. Crushed ore is piled on a plastic-lined pad, sprinkled with cyanide solution, and the pregnant solution drains to a recovery circuit. Capital cost is a fraction of CIL, but recoveries are lower (60–75% vs 90%+) and the cycle takes weeks to months. This is the technology that unlocked low-grade Nevada and Australian oxide ore in the 1980s–90s and made super-majors of Newmont (Carlin Trend) and Barrick (Goldstrike).
• Refractory processing — where gold is locked inside sulfide crystals, the sulfides must be oxidized first via roasting, pressure oxidation (autoclaves), or biological oxidation. This is expensive and is why some "high-grade" deposits never get built.
• Pressure oxidation (autoclave) — the highest-capex, highest-opex route, used for high-sulfide refractory ore, achieving 90%+ recovery.

Stage 5 — Refining (the LBMA refiners). The 80–90% gold doré must be refined to 99.5% (Good Delivery bullion bar) or 99.99% (4-nines investment bar). Refiners use the Miller process (chlorine gas removes base metals, gets to 99.5%) or the Wohlwill process (electrolysis, gets to 99.99%). The LBMA Good Delivery List is the ticket to global markets — only listed refiners produce bars accepted in the London OTC market without re-assay. To get on the list a refiner must produce at least 10 tonnes of refined gold per year, hold £15 million tangible net worth, have operated at least 5 years, and pass an annual responsible-sourcing audit. There are only about 60 LBMA-accredited gold refiners globally: Valcambi, Argor-Heraeus, Metalor, PAMP (all Swiss), Rand Refinery (South Africa), Royal Canadian Mint, Perth Mint, Tanaka (Japan). Switzerland alone refines roughly two-thirds of global newly-mined gold — a moat built by Swiss banks in the 1960s and hard to replicate because of regulatory and trust infrastructure.

Stage 6 — The market (vaults, banks, end users). Refined bars flow into three buckets: (1) the LBMA vault network (London plus Zurich, New York, Singapore), where bullion-bank inventory and ETF holdings sit and the London market clears about $200 billion of paper gold per day on top of physical settlement; (2) central bank reserves, around 36,000 tonnes globally (US 8,133t, Germany 3,355t, Russia 2,336t, China 2,298t acknowledged and likely higher); (3) jewelry and industrial demand, roughly 2,200–2,500 tonnes per year with India and China over half, plus small but inelastic industrial demand (electronics bonding wire, dental, catalyst).

Mining-method tradeoffs — geology dictates everything

There is no universally best method. The geology fixes the choice, and the choice cascades into capital intensity, mine life, recovery, and labor profile.

Approach	When you use it	Capex	Opex	Recovery	Suppliers
Open pit + heap leach	Shallow, low-grade, oxide ores	Lowest	Lowest (sub-$900/oz possible)	60–75%	Cat, Komatsu, drilling consumables
Open pit + CIL mill	Shallow, low-to-mid grade, free-milling	Mid	Mid ($1,000–1,400/oz)	90–95%	Cat, Komatsu, Outotec/Metso for mill
Underground + CIL mill	Deep, high-grade veins	Highest	Variable ($800–1,800/oz)	90–95%	Sandvik, Epiroc, Cat, Outotec
Refractory (POX/roast)	Sulfide-locked gold	Very high	High	85–92%	Outotec, FLSmidth
Pressure oxidation (autoclave)	High-sulfide refractory	Highest	Highest	90%+	Specialty engineering

A company running multiple deposits across multiple methods (Newmont, Barrick, AEM) gets diversification but loses focus. A company running one or two assets on the same method (WDO, LUG) gets focus but takes single-asset risk. The high-grade underground model is a fundamentally different economic animal from low-grade open pit: less dirt to move per ounce, a smaller environmental footprint, lower energy intensity, but higher technical complexity. WDO runs the highest reserve grade in the listed universe at 12.67 g/t, and LUG's Fruta del Norte and AGI's Island Gold both run 10+ g/t — among the highest-grade large gold mines in the world.

The unit economics — where margin and operating leverage come from

Producers have zero pricing power on the gold itself (a global, fungible commodity priced by COMEX, LBMA, and the Shanghai Gold Exchange). Their only pricing power is on the cost side (negotiating with equipment vendors and labor) and in capital allocation. That structure makes them a leveraged play on the gold price: because AISC is a roughly fixed dollar cost per ounce, every dollar of gold price above AISC drops to cash margin, so a percentage move in gold becomes a larger percentage move in cash flow. WDO is the cleanest illustration — its operating leverage to gold is quantified at ~2.5x. The cash margin per ounce at $4,700 spot ranged across the screened producers from $3,036/oz (EGO) to $3,685/oz (LUG), with the spread driven almost entirely by AISC differences (EGO $1,664 vs LUG $1,015).

The capital cycle compounds the leverage. Gold mining runs two cycles at different timescales: the gold-price cycle (5–7 years between major peaks) and the capital cycle (a 7–10 year lag between price highs and new mine completions). When both align bullishly — high price, no new supply for years — operating margins hit record highs, which is the 2025–2026 setup.

Brownfield expansion economics deserve their own note because they are the highest-return capital deployment in the sector. A producer extending an existing permitted, high-grade orebody avoids the 10–15-year permitting slog and the $1.5–2.5B greenfield build. AGI's Island Gold Phase 3+ shaft is the textbook case: the project NPV is $8.16B with a 53% IRR at $3,200 gold (and 69% IRR / NPV $12.2B at $4,500 gold), with project-level break-even sub-$1,200/oz, taking the mine from ~130koz to 287koz annually at sub-$900/oz, and group mine-site AISC toward sub-$1,025/oz. The reason the economics work: the long-lead infrastructure is sunk — the shaft was reported sunk to 1,350 m of 1,379 m planned, with headframe, hoist house, bin house, and paste plant in late-stage commissioning, and 91% of the total growth-capital commitment already spent. That converts a "leveraged gold bet" into a brownfield-expansion compounder whose returns do not require gold to keep rising.

The cost-structure drivers — and why AISC is creeping

AISC for the industry rose from roughly $1,000 in 2020 to ~$1,521 in 2025, even as gold ripped. The drivers are structural and worth separating:

• Energy — the mill (SAG and ball mills) is the heaviest user; in deep underground mines, ventilation is the largest single opex line, which is why battery-electric underground equipment (removing diesel removes ventilation load) is the one technology genuinely changing the cost curve.
• Labor, equipment, and grinding-media inflation — all up across the cycle; mining tires are a chronic supply bottleneck and a high-margin consumable.
• Royalties — government take is rising in every jurisdiction. The 2025 AISC creep at AEM (about 12%) was attributed by management as "more than half" cyclical — royalty step-ups (which scale with the gold price) plus CAD strength against a USD-denominated gold price.
• Reagents — sodium cyanide (the leach reagent) is a quiet Orica/Cyanco duopoly; activated carbon wears out and recurs; lime and flotation reagents recur with throughput.

The geological headwind — gold's Moore's Law runs backwards

The productivity trend in gold mining moves the wrong way. Average reserve grade for major producers has fallen from roughly 1.5 g/t in 2000 to about 1.0 g/t today. No major deposit (over 2 million ounces) has been found since 2022; reserves at the top 20 producers fell 26% from 2012–17 and have not fully recovered. New mine builds take 10–15 years from discovery to first pour. Mine production hit a record 3,672 tonnes in 2025 yet barely keeps up with demand because the industry is running into peak-gold geology. The honest summary: gold mining is a slowly improving heavy industry with productivity gains of 1–2% per year — no new science has changed the economics since heap leaching in the 1980s. AI-driven exploration (KoBold, IBM) is a real productivity boost but does not create new ore bodies; autonomous trucks, battery-electric LHDs, in-situ leaching (works for uranium/copper, not yet gold), and tailings reprocessing are incremental. The investment case is the price-margin spread on mostly the same operations, not technology disruption.

Where the engineering creates moats

The moats are not in the metal — they are in the "hard" steps and in permitted ore. Permitting a new mine takes 7–12 years in Nevada or Quebec and 15+ in Chile, Peru, or the US; anyone holding permitted ore in the ground at $1,500 AISC has an essentially uncopyable moat for a decade. Reserve life is the durability metric: below 10 years is short, above 15 is comfortable. WDO's ~6-year reserve life (1.13 Moz P+P against ~180koz/yr) means it must drill its way out of depletion every year, while AEM's ~16 years and AGI's ~25 years post-Phase-3+ are the structural cushion. Reserve replacement — growing reserves faster than you mine them — is the rarest skill; AEM grew reserves 2% in 2025 to 55.4 Moz while mining 3.4 Moz, the case where exploration genuinely replaces depletion.

The parallel royalty/streaming chain monetizes a different moat: capital and deal flow rather than rock. A royalty/streaming company (FNV, Wheaton, Royal Gold) pays miners upfront cash for a fixed percentage of future production or the right to buy gold at a fixed price, carrying zero operational, capex, or permitting risk. The economics are the cleanest in the entire chain: 70–80% gross margins (FNV runs ~91% EBITDA margin and ~82% FCF margin), near-1:1 operating leverage to the gold price, and NAV-per-share compounding of 10–12% annually for two decades. The top three control roughly 80% of contract value. The one risk the model does not escape is political risk transmitted through the underlying mines — FNV's Cobre Panama royalty, shut by the Panamanian government in 2023, cost it ~15% of revenue, the lesson that even royalty companies inherit jurisdictional risk through their assets.

Where the profit pools sit across the chain

Layer	Approx revenue pool (2026E)	Gross margin	Concentration	Barriers to entry
Exploration (juniors)	$5–8B	Negative	Highly fragmented	Geological skill, capital
Mining equipment OEMs	$90–100B	30–40%	Oligopoly (Cat/Komatsu/Sandvik/Epiroc)	Scale, dealer network, IP
Engineering & construction (EPCM)	$20–30B	5–10%	Fragmented	Reputation, track record
Producers (the miners)	$200B+	40–60% at current gold price	Top 10 = 30% of supply	Reserves, permits, jurisdictions
Refiners (LBMA)	$1–2B in fees	10–20%	Switzerland-dominated	LBMA accreditation, bank relationships
Bullion banks (LBMA market makers)	$5–10B	High	Top 5 dominate	Capital, balance sheet, regulatory
Royalty & streaming	$4–6B	70–80%	Top 3 = 80%	Upfront capital, deal flow
Vaulting & custody	$1–2B	40–50%	Concentrated	Trust, infrastructure

The producer side is more fragmented than people assume — the top 10 producers account for only ~30% of global mine supply, and the biggest single producer (Newmont) has about 6% global share. There is no Aramco or Vale equivalent in gold, partly because deposits are geographically scattered and partly because national governments insist on local ownership. The picks-and-shovels are the opposite: Cat, Komatsu, Hitachi, and Liebherr hold 80%+ of global surface mining truck share; Sandvik, Epiroc, and Cat ~75% of underground equipment. The reason for the oligopoly is brutal capital intensity, decade-long product cycles, and a global dealer/service network that takes 50 years to build. Global gold mine production of 3,672 tonnes in 2025 was worth ~$390 billion at the $3,300 average; at $4,700 spot the run-rate is closer to $550 billion of producer revenue, with EBITDA margins of 25–50% (40–60% at current prices) and reserves, permits, and jurisdiction quality as the durable barriers.

Subsectors

The gold and precious-metals mining sector is not one business. It is a stack of distinct sub-areas, each with its own economics, technology, risk profile, and reason to own it. The producers sit at the center, but they split sharply by scale and lifecycle stage, and the most attractive economics in the whole chain sit one layer removed in royalty and streaming. Two analytical lenses — jurisdiction screening and mining economics (AISC) — cut across every producer and decide which names are actually ownable. This section enumerates each.

Senior producers

The seniors are the large, diversified, multi-mine companies at the top of the production table. The defining trait is scale and diversification: 6+ mines across multiple countries, production measured in millions of ounces per year, and no single asset large enough to sink the company. Newmont (NEM) is the biggest single producer globally at ~6.0 Moz in 2025 and ~6% global share — and even the largest player has only single-digit share, because the producer side is far more fragmented than people assume (top 10 producers ≈ 30% of global mine supply, no Saudi Aramco or Vale equivalent in gold).

Who plays: Newmont (NEM, ~6.0 Moz, 135 Moz reserves, $80-90B cap, Africa-exposed via Ahafo/Akyem), Barrick Mining Corp (B/GOLD, 3.9 Moz, 89 Moz reserves, Africa via Kibali/Loulo/Tongon), Agnico Eagle (AEM, 3,450 koz / 3.4 Moz, 55.4 Moz reserves, $80-90B+ cap, no Africa), AngloGold Ashanti (AU, 2.6 Moz, Africa-exposed), Gold Fields (GFI, 2.2 Moz, South Africa HQ, Africa-exposed), Kinross (KGC, 2.1 Moz, Tasiast/Mauritania = 25% of production), and Northern Star (NST.AX, 1.8 Moz, Australian-listed, no Africa).

The technology is diversification of method: a senior typically runs open-pit + heap leach, open-pit + CIL mill, and underground + CIL mill across its portfolio simultaneously. That buys resilience but loses focus — a Newmont/Barrick/Agnico gets method and jurisdiction diversification at the cost of being unable to move the needle on any single mine. Reserve life tends to be comfortable: AEM carries ~16 years across the portfolio, which is rare.

The investment angle is "own the cycle." Seniors are the lowest-volatility way to hold producer exposure; the quality compounders among them (Agnico the textbook example) combine Tier-1 jurisdictions, low-quartile AISC, disciplined capital allocation, and reserve replacement through exploration rather than depletion. AEM grew reserves 2% in 2025 to 55.4 Moz despite mining 3.4 Moz — the rare case where exploration replaces depletion. The catch is valuation: quality seniors trade at a P/NAV premium (AEM 1.4-1.6x vs sector 1.0-1.2x), and if gold pulls back they compress faster than discount-priced miners. The bifurcation among seniors is between the disciplined operators (Agnico, who bought back stock at $42 in 2022 and returned $1.4B in 2025) and the M&A junkies (Barrick under various CEOs) — the 2011-2015 bear market killed the latter when companies that paid up at $1,800 gold ate billion-dollar writedowns at $1,050. Investors still discount the undisciplined to mid-single-digit P/E because they don't trust them to spend the windfall well.

Mid-tier / growth producers

The mid-tiers are the sweet-spot names: big enough to be in indices and have institutional liquidity, small enough that a single brownfield expansion or new mine moves EPS materially. This is where production growth lives. In the no-Africa Canadian universe, AGI and EGO are the only two names with material growth profiles; the seniors are flat-to-gentle and the small-caps are stable.

Who plays (Canadian no-Africa cut): Alamos Gold (AGI, 545,400 oz 2025, targeting 900,000+ by 2028 = +65%, 15.9 Moz reserves +32% YoY on Magino, $19.4B cap, net cash, pure North America), Eldorado Gold (EGO, 488,268 oz 2025, targeting 620,000-720,000 by 2027 = +40%, 12.5 Moz reserves, $9.4B cap), Lundin Gold (LUG, 498,315 oz, single-asset Fruta del Norte in Ecuador). Broader mid-tier set includes Endeavour Mining (EDV, pure West Africa), Pan American Silver (PAAS), B2Gold (BTG, Mali/Namibia), and IAMGOLD (IAG, Burkina Faso/Mali).

The technology angle for mid-tiers is usually a single high-conviction project: AGI's Island Gold Phase 3+ (one of the highest-grade gold mines in the Western world at 10+ g/t, shaft sunk to 1,350 of 1,379 m planned, 91-98% of growth capex spent), or EGO's Skouries copper-gold porphyry in northern Greece (sub-$700/oz net of copper credits, adds ~140 koz gold + ~67 Mlbs copper, ramping 2026 after 15+ years of permitting fights). The growth is mechanically driven by de-risked infrastructure, not commodity price.

The investment angle is GARP — growth at a reasonable price, with re-rating optionality. The thesis is "buy tomorrow's senior at today's mid-tier multiple": AGI at ~22x trailing on a name about to grow earnings 65% over three years, re-rating to senior multiples on execution. AGI is the highest-conviction name in the whole screen specifically because its growth doesn't require gold to keep ripping — the Phase 3+ NPV works at $3,200 gold (Alamos's base case), NPV $8.16B, 53% IRR, project break-even sub-$1,200/oz, and 2026 is the FCF trough (capex peak $895M, FCF compresses to $50-100M) before FCF returns to ~$1.1B by 2028 (6% FCF yield on the $19B cap). The risk is execution: a slipped expansion or a depleting legacy mine (AGI's Mulatos) and the growth disappears. Mid-tiers also carry less liquidity in stress and smaller scale than seniors.

Royalty & streaming

This is the parallel chain, and it carries the most attractive economics in the entire sector — highest margin, lowest operational risk. A royalty/streaming company pays miners upfront cash in exchange for either a fixed percentage of future production (royalty) or the right to buy metal at a fixed low price (stream). It owns gold-price exposure without owning the mine: zero capex risk, zero permitting risk, zero operational risk, near-1:1 operating leverage to the gold price.

The profit-pool math: gross margins of 70-80% across the layer, a revenue pool of $4-6B (2026E), and a top-3 concentration of ~80% of contract value. The model is a structural oligopoly — barriers to entry are high because you need the relationships, the deal flow, and a balance sheet large enough to write checks competitive with Franco-Nevada.

Who plays: Franco-Nevada (FNV, $1.2B 2025 revenue, $35-40B cap, the largest and most diversified — 430 assets / 119 producing across gold/silver/base metals/oil and gas, no single asset >10-13% of revenue, 91% EBITDA margin, 82% FCF margin, zero debt, $3.1B available capital, 19 consecutive annual dividend hikes, NAV/share compounding 10-12% annually for two decades — "the Berkshire of gold"); Wheaton Precious Metals (WPM, ~$1.0B revenue, $30-35B cap, streaming-focused, gold + silver, growth-tilted, more leveraged to PM prices); Royal Gold (RGLD, ~$700M revenue, $12-15B cap, smallest of the big three, US-listed, more concentrated); Triple Flag (TFPM, ~$200M, mid-cap, growing); Sandstorm Gold (SAND, ~$200M, junior-focused, more aggressive/higher beta); Osisko Gold Royalties (OR, ~$250M, tied to specific mid-tier producers). Junior streamers — Metalla, Vox Royalty, Empress Royalty — are higher-risk, higher-beta versions of the FNV model.

The investment angle is ballast and compounding. Streamers compound mid-teens through cycles with much less drawdown risk; the rule of thumb is streamers compound mid-teens through cycles, miners compound high-teens-to-twenties in good cycles and zero in bad ones. The textbook structure is core-and-satellite: FNV as the ballast core (strips out operational risk, compounds steadily) paired with a producer satellite for upside. The risk is twofold — premium valuation (FNV P/NAV 1.8-2.2x, ~28-30x P/E priced for continuation, compresses fastest if gold stalls) and the lesson that even royalty companies carry political risk through their underlying mines: Cobre Panama (a major FNV royalty asset shut down by Panama's government in 2023) cost FNV ~15% of revenue, is treated as zero in 2026 guidance, and resolution (ICC arbitration hard date October 2026) is free upside optionality rather than base case.

Jurisdiction screening

Jurisdiction is not a sub-business but it is a distinct analytical sub-area — the screen that decides which producers are ownable at all, and it is "the whole game." The reference framework is the Fraser Institute Investment Attractiveness Index, ranked annually. Tier 1: Canada (Quebec, Ontario, Yukon, Nunavut), Australia (Western Australia), Nevada, Finland, Sweden. Middle tier: Mexico, Chile, Brazil (parts), Türkiye, Greece, Ecuador (variable). Bottom tier: Mali, DRC, Burkina Faso, Tanzania, Russia, Venezuela, Indonesia (some areas).

What changed the math: country risk used to discount valuations 10-20%; in 2026 it discounts them 30-50%. West African gold mines have been seized by junta governments in Burkina Faso, Mali, and Niger across 2022-2025. This is the structural basis for the "Doug filter" — Canadian-listed, no Africa exposure — which screens out Barrick (Africa), Kinross (Tasiast/Mauritania = 25% of production), B2Gold (Mali Fekola), and IAMGOLD (Burkina Faso). The five names that survive the filter: AEM, AGI, EGO, WDO, LUG.

Royalty regimes are the second jurisdictional axis — government take varies wildly and the post-2022 trend everywhere is "more for the state": Quebec ~16% effective tax + provincial royalty; Nevada ~5% net proceeds tax; Chile ~5% gross royalty plus corporate tax; Mali 6% royalty + 35% tax + 10% state free-carry on new mines. Even Tier-1 Canada raised provincial mining taxes in 2024-25.

The investment angle: permitting scarcity is the underappreciated moat. A new mine takes 7-12 years from discovery to first pour in Nevada or Quebec, 15+ in Chile/Peru/US — so anyone holding permitted ore in the ground at ~$1,500 AISC in a Tier-1 jurisdiction has an essentially uncopyable advantage for a decade. The structural disruption pattern is "permitted ore in a Tier-1 jurisdiction taken from a junior to mid-tier producer status" (Skeena's Eskay Creek redevelopment in BC, G Mining's Tocantinzinho in Brazil are the cleanest examples). The jurisdictional ranking of the no-Africa five runs WDO/AEM (lowest) through EGO/LUG (highest, on Türkiye ~50% and single-country Ecuador respectively).

Mining economics / AISC

AISC (all-in sustaining cost, $/oz) is the cross-cutting metric that determines whether any producer's stock works, and it deserves its own sub-area because it drives the cash-margin math behind the whole sector thesis. AISC is the all-in cash cost to produce one ounce including direct mining, processing, royalties, sustaining capex, and G&A, but excluding growth capex. The industry average was $1,521/oz in 2025 (up from ~$1,000 in 2020). Reading thresholds: below $1,200 is excellent in 2026, $1,200-1,500 acceptable, above $1,800 means high-cost and only works at high gold prices.

The cash-margin engine is what makes the sector compelling: at ~$4,700 spot and ~$1,500 AISC, producers run ~$3,200/oz gross cash margin, and World Gold Council math implies industry margin of roughly $2,800/oz in 2026 — the highest dollar-margin per ounce in history. AISC margin (gold price minus AISC) is the single number that converts the gold price into producer earnings.

The economics are dictated by mining method, which the geology chooses:

Method	When used	Capex	Opex ($/oz)	Recovery
Open pit + heap leach	Shallow, low-grade, oxide	Lowest	sub-$900	60-75%
Open pit + CIL mill	Shallow, low-to-mid grade, free-milling	Mid	$1,000-1,400	90-95%
Underground + CIL mill	Deep, high-grade veins	Highest	$800-1,800 (variable)	90-95%
Refractory (POX/roast)	Sulfide-locked gold	Very high	High	85-92%
Pressure oxidation (autoclave)	High-sulfide refractory	Highest	Highest	90%+

The no-Africa five span the cost curve sharply. Per-ounce 2025 AISC and the cash margin each captures at $4,700 spot:

	AEM	AGI	EGO	WDO	LUG
2025 AISC ($/oz)	$1,339	$1,524	$1,664	$1,518	$1,015
Cash margin/oz at $4,700	$3,361	$3,176	$3,036	$3,182	$3,685

LUG is the cost king globally at $1,015/oz — over $300 below the next-best and ~$649 below the highest (EGO), worth roughly $150-200M of extra free cash per year vs comparable peers. AGI is the only name that gets cheaper through 2028 (Island Gold Phase 3+ takes mine-site AISC to sub-$1,025/oz on the largest production block) while peers see cost creep. AEM's 2025 AISC sits in the lowest quartile among seniors but its 2026 guide implies ~12% creep — flagged as "mostly cyclical" (royalty step-ups + CAD strength, "more than half" per CEO commentary). EGO's group AISC is misleading because Skouries runs at negative AISC net of copper credits, so 2027 pro-forma cost looks very different from 2026.

Two adjacent economic concepts matter for the AISC sub-area. First, reserve pricing: reserves are calculated at a conservative gold-price assumption (AEM books at $1,600/oz, only ~35% of spot), so M&I + Inferred ounces (AEM 88.9 Moz combined) are huge upside optionality at spot. Second, the structural cost headwind — "the Moore's Law equivalent for gold mining is moving in the wrong direction": average reserve grade for major producers fell from ~1.5 g/t in 2000 to ~1.0 g/t today, deposits are getting deeper, lower-grade, and harder to permit. The investment angle is that AISC discipline is the dividing line between the names you own (low-quartile, stable or falling AISC) and the names you avoid (high-cost, only work at elevated gold). The operating-leverage corollary: high-grade underground pure-plays like WDO (12.67 g/t reserve grade, highest in the listed universe) carry ~2.5x operating leverage to the gold price, the highest beta in the screen.

Small-cap / high-grade pure-plays

Distinct from the mid-tiers is a small-cap sub-area defined by single-jurisdiction, high-grade, high-beta exposure. WDO (Wesdome) is the archetype: two underground mines, both Canadian (Eagle River in Wawa, Ontario; Kiena in Val-d'Or, Quebec), 100% Canada, 100% underground, 100% high-grade, with the highest reserve grade in the listed universe (12.67 g/t). Production is ~186 koz (2025), reserves only 1.13 Moz P+P — a ~6-year reserve life that is the defining risk, requiring continuous exploration success to extend. K92 Mining (KNT, Papua New Guinea high-grade) and Centerra (CGAU, Türkiye + Canada) sit in the same scale bucket.

The technology is high-grade underground: less dirt moved per ounce, smaller footprint, lower energy intensity, but higher technical complexity (WDO's Q4 2025 AISC spiked to $1,750 on the Kiena hoist shutdown). The investment angle is pure operating leverage to gold — ~2.5x — making it the highest-beta way to play a continued rally, but only as a basket sleeve rather than a single-stock bet given the short reserve life and two-mine concentration (a geotechnical event at either site cuts production 40-50%). The June 2026 NI 43-101 reserve update is the swing data point on which the whole reserve-replacement thesis stands or falls.

Adjacent / emerging sub-areas

Two further sub-areas sit at the edges. The first is exploration juniors — the lottery-ticket layer ($5-8B pool, negative gross margins, highly fragmented; roughly 1 in 1,000 explored targets becomes a producing mine). The interesting pattern is "permitted ore in a Tier-1 jurisdiction moving from junior to mid-tier" rather than wildcat discovery: Skeena Resources (SKE, Eskay Creek redevelopment in BC, first-pour 2027), G Mining Ventures (GMIN, Tocantinzinho Brazil), Calibre Mining (CXB, Nicaragua + Nevada), Snowline Gold (SGD, Yukon Rogue discovery). The second is copper-gold porphyry convergence — deposits built primarily for copper that throw off significant gold credits (Skouries, Cascabel in Ecuador, Filo Mining's Filo del Sol in Argentina), which is how much incremental gold supply will come online over the next decade. Both are watchlist-grade for this filter rather than core positions.

ETF baskets give pooled exposure across these sub-areas: GDX (VanEck Gold Miners, senior + mid-tier, the default basket), GDXJ (Junior Gold Miners, more leverage/volatility), SGDM (Sprott, quality-weighted to strong balance sheets), and GOAU (US Global, heavy royalty/streamer weighting — the cleanest "premium quality" basket).

Value chain

The gold value chain is six distinct stages from rock in the ground to a delivered Good Delivery bar in an LBMA vault. Each stage has its own economics, its own technology, its own suppliers — and each "hard" step (hard to find, hard to extract economically, hard to refine to bullion purity, hard to move and store) is a separate profit pool. The single most important fact about the chain: producers have zero pricing power on gold itself. Gold is a global, fungible commodity. The only pricing power producers hold is on the cost side (negotiating with equipment vendors and labor) and on capital allocation (what to do with windfall margins). Everything below maps where the margin actually sits.

The six stages, end to end

Stage 1 — Exploration (5 to 15 years, junior miners). Geologists identify targets through remote sensing, geochemical sampling (panning streams, soil grids), and geophysics (magnetic, gravity, induced-polarization surveys), then drill thousands of meters of diamond core to map the ore body in 3D. A "discovery" usually means a few hundred thousand ounces inferred; a "deposit worth building" means a few million ounces with grade and metallurgy that pencil at conservative gold prices. Roughly 1 in 1,000 explored targets becomes a producing mine. Economics: ~$5-8B global revenue pool, negative gross margin, highly fragmented, value is optionality. These are lottery tickets.

Stage 2 — Development (3 to 7 years, $500M to $2B+ capex). Once a deposit is delineated and permitted — the slow part: environmental review, indigenous consultation, water rights, tailings approvals can take a decade — construction begins. Stripping overburden in open pits, sinking shafts and developing levels underground, building the mill, the tailings dam, the camp, the power line. Capital intensity has gotten worse: a new million-ounce-per-year project today costs $1.5-2.5 billion. This is precisely why seniors don't build greenfield and instead acquire juniors that already hold permitted deposits. EPCM (engineering, procurement, construction management) contractors play here: ~$20-30B pool, but thin 5-10% gross margin, fragmented, commoditized.

Stage 3 — Mining (the operating mine). Two approaches, dictated by geology:

• Open pit: bench-by-bench excavation of shallow, low-grade, large-tonnage deposits. Massive haul trucks (Caterpillar 797F at 400-tonne capacity), electric rope shovels, bulldozers. Strip ratio (waste tonnes moved per ore tonne) drives economics. Caterpillar, Komatsu, Hitachi, Liebherr sell most of the gear here.
• Underground: shafts, declines (spiral ramps for trucks), levels, stopes. Trackless equipment — load-haul-dump (LHD) machines, jumbo drill rigs, articulated trucks — moves ore to a hoist or up a decline. Sandvik and Epiroc territory: those two plus Caterpillar are ~75% of the global underground equipment fleet.

Stage 4 — Milling and processing (the plant). This is where the chemistry happens and where the heaviest energy load sits. Ore is crushed (jaw then cone crusher), ground (semi-autogenous and ball mills — the most energy-intensive step in the whole flowsheet), then leached. The dominant leach for free-milling ore is CIL (carbon-in-leach):

Crushed ore → Ball mill → Cyanide leach tank (with carbon)
                              ↓
                    Loaded carbon (carries gold)
                              ↓
                    Strip vessel (high T, high P)
                              ↓
                    Gold-bearing electrolyte
                              ↓
                    Electrowinning cells
                              ↓
                    Gold sludge → Smelt → Doré bar

Cyanide dissolves gold into solution; activated carbon in the same tank captures it like a sponge; loaded carbon is washed with hot caustic to release gold into concentrated electrolyte; electric current at the electrowinning cell deposits gold as metallic sludge on cathodes; the sludge is melted in an induction furnace, and out comes the doré bar (roughly 80-90% gold). For lower-grade or oxide ore, heap leach is often cheaper — crushed ore piled on a plastic-lined pad, sprinkled with cyanide solution, pregnant solution drained to a recovery circuit. Capital cost is a fraction of CIL but recoveries are lower (60-75% vs 90%+) and the cycle runs weeks to months. For refractory ore (gold locked inside sulfide crystals), you must oxidize the sulfides first via roasting, pressure oxidation (autoclaves), or biological oxidation — expensive, and the reason some "high-grade" deposits never get built.

Stage 5 — Refining (the LBMA refiners). The doré bar leaving the mine site is ~80-90% gold; to trade on world markets it must be refined to 99.5% (Good Delivery bullion bar) or 99.99% (4-nines investment bar), using the Miller process (chlorine gas, gets to 99.5%) or the Wohlwill process (electrolysis, gets to 99.99%). This is a severe choke point. The LBMA Good Delivery List is the ticket to global markets — only listed refiners can produce bars accepted in the London OTC market without re-assay. To get on the list a refiner must produce at least 10 tonnes of refined gold per year, hold £15 million tangible net worth, have operated at least 5 years, and pass an annual responsible-sourcing audit. There are only about 60 LBMA-accredited gold refiners globally: Valcambi, Argor-Heraeus, Metalor, PAMP (all Swiss), Rand Refinery (South Africa), Royal Canadian Mint, Perth Mint, Tanaka (Japan). Switzerland alone refines roughly two-thirds of global newly-mined gold — because Swiss banks built the trade in the 1960s and the regulatory and trust infrastructure is hard to replicate. Refiner economics are thin in dollar terms (process margin $5-15/oz, ~$1-2B in total fees, 10-20% gross margin) but the position is structurally protected by accreditation and bank relationships.

Stage 6 — The market (vaults, banks, end users). Refined bars flow into three buckets:

1. LBMA vault network (London, plus Zurich, New York, Singapore) — where bullion-bank inventory and ETF holdings sit. London clears about $200 billion of paper gold per day on top of physical settlement.
2. Central bank reserves, ~36,000 tonnes globally. US 8,133t, Germany 3,355t, Russia 2,336t, China 2,298t (acknowledged; likely higher). China's PBoC bought every month for 15+ consecutive months as of early 2026.
3. Jewelry and industrial demand, ~2,200-2,500 tonnes/year. India and China together over half. Industrial (electronics bonding wire, dental, catalyst) is small but inelastic.

The value-chain map and where the profit sits

[EXPLORATION]                  [DEVELOPMENT]                [MINING & PROCESSING]
Junior explorers      →        Permitting & build      →   Producers (the miners)
~$5B/yr global           ~$10B/yr capex globally          ~$200B/yr revenue at $4,700/oz
Very low margins,            Negative cash flow              25-50% EBITDA margins,
optionality value             until first pour               cyclical to gold price
                                                                      ↓
[REFINING]                ←   Doré (80-90% pure)
~60 LBMA refiners
Process margin: $5-15/oz
Highly concentrated (Switzerland)
        ↓
[VAULTING & MARKET]
LBMA bullion banks (JP Morgan, HSBC, ICBC Standard, UBS, etc.)
COMEX, SGE, LBMA
Custody, lending, derivatives
        ↓
[END USE]
Central banks (~25% of demand)
Jewelry (~50%, India + China dominant)
ETFs and bars/coins (~20%, swing factor)
Industrial (~5%, electronics, dental)

[PARALLEL CHAIN: ROYALTY/STREAMING]
Franco-Nevada, Wheaton, Royal Gold provide upfront capital to miners
in exchange for royalties on production or rights to buy gold at fixed prices.
Top 3 control ~80% of contract value.

The profit-pool table is the load-bearing piece of the chain. It tells you where margin pools and where it doesn't:

Layer	Approx revenue pool (2026E)	Gross margin	Concentration	Barriers to entry
Exploration (juniors)	$5-8B	Negative	Highly fragmented	Geological skill, capital
Mining equipment OEMs	$90-100B	30-40%	Oligopoly (Cat/Komatsu/Sandvik/Epiroc)	Scale, dealer network, IP
Engineering & construction (EPCM)	$20-30B	5-10%	Fragmented	Reputation, project track record
Producers (the miners)	$200B+	40-60% at current gold price	Top 10 = 30% of supply	Reserves, permits, jurisdictions
Refiners (LBMA)	$1-2B in fees	10-20%	Switzerland-dominated	LBMA accreditation, bank relationships
Bullion banks (LBMA market makers)	$5-10B	High	Top 5 dominate	Capital, balance sheet, regulatory
Royalty & streaming	$4-6B	70-80%	Top 3 = 80%	Upfront capital, deal flow
Vaulting & custody	$1-2B	40-50%	Concentrated	Trust, infrastructure

The most attractive economics, layer by layer: royalty and streaming (highest margin, lowest operational risk), producers (the leveraged play on gold price, but the only stage exposed to operational and cost risk), and the equipment oligopoly (steady, secular). The least attractive: exploration juniors (lottery tickets) and EPCM contractors (commoditized).

The bottleneck tiers — who actually controls the chain

Three layers are tight oligopolies or near-monopolies; the producer layer itself is the fragmented one. This inversion is the central structural fact of the sector.

Producers are fragmented at the top. Top 10 producers are only ~30% of global mine supply. There is no Saudi Aramco or Vale equivalent in gold — the biggest single producer (Newmont) has ~6% global share. Deposits are geographically scattered and national governments insist on local ownership of strategic resources. Top 10 = 30% of supply.

Picks-and-shovels are an oligopoly. Mining equipment is the opposite of the producer layer. Cat, Komatsu, Hitachi, Liebherr hold over 80% of global surface mining truck market share. Sandvik, Epiroc, and Cat hold ~75% of underground equipment and ~88% of LHDs. The moat is brutal capital intensity, decades-long product-development cycles, and a global dealer/service network that takes 50 years to build. Nobody disrupts these from a garage.

Reagents are a quiet duopoly. Sodium cyanide — the gold-leaching reagent every CIL/heap-leach mine consumes — comes from Orica (ASX: ORI) and Cyanco (Draslovka, private), who supply most global gold production. Recurring, consumable, supply-constrained.

Mining tires are a chronic bottleneck. Bridgestone, Michelin, Goodyear. Mining tires are described in the sources as a chronic supply bottleneck and high-margin business — a recurring physical choke point on haul-truck uptime.

Refining is geographically concentrated to the point of fragility. ~60 LBMA refiners worldwide, Switzerland refining two-thirds of newly-mined gold. The accreditation requirements (10 tonnes/year minimum, £15M net worth, 5-year track record, annual audit) function as the entry barrier.

Royalty/streaming is a top-3 game. Franco-Nevada, Wheaton, Royal Gold control ~80% of contract value. Barrier is upfront capital and deal flow.

Bullion banks: top 5 dominate the LBMA market-making layer (JP Morgan, HSBC, ICBC Standard, UBS, Goldman, Morgan Stanley, BNP Paribas), gated by balance sheet and regulatory capital.

Supplier and consumable relationships by stage

The supplier map matters because it tells you who recurs with production (and is therefore investable on the same secular tailwind) versus who's cyclical to exploration spend:

• Exploration services: Diamond drilling (Major Drilling / TSX: MDI, Boart Longyear, Foraco) — cyclical, leveraged to junior exploration spend. Assay labs (SGS, ALS / ASX: ALQ, Bureau Veritas). 3D modeling software (Dassault GEOVIA, Seequent/Bentley, Maptek) — subscription, sticky.
• Surface mining: Caterpillar (CAT), Komatsu (6301), Hitachi (6305), Liebherr.
• Underground mining: Sandvik (SAND), Epiroc (EPI), Caterpillar.
• Drilling consumables: Sandvik drill bits, Epiroc, Atlas Copco, Robit — wear out fast, recur with production.
• Mill liners and grinding media: Magotteaux, Moly-Cop, AIA Engineering (NSE: AIAENG, world leader in high-chrome grinding media).
• Mill equipment (SAG/ball mills, crushers): Outotec/Metso (HEL: METSO), FLSmidth (FLS), Weir Group (WEIR) — the Big Three of mineral-processing equipment.
• Pumps and slurry: Weir, Schlumberger (Cameron), Sulzer.
• Reagents: Sodium cyanide (Orica, Cyanco). Activated carbon (Cabot, Calgon/Kuraray, Haycarb) — wears out, recurs. Lime and flotation reagents (BASF, Solvay, Nouryon).
• Price discovery / venue: CME Group (COMEX gold futures — the price-discovery venue), HKEX, Shanghai Gold Exchange.

Per-stage and per-method economics

Mining method is not a choice; geology dictates it, and the choice cascades into capital intensity, mine life, and labor profile. The tradeoff table:

Approach	When you use it	Capex	Opex	Recovery	Suppliers
Open pit + heap leach	Shallow, low-grade, oxide ores	Lowest	Lowest ($/oz can be sub-$900)	60-75%	Cat, Komatsu, drilling consumables
Open pit + CIL mill	Shallow, low-to-mid grade, free-milling	Mid	Mid ($1,000-1,400/oz)	90-95%	Cat, Komatsu, Outotec/Metso for mill
Underground + CIL mill	Deep, high-grade veins	Highest	Variable ($800-1,800/oz)	90-95%	Sandvik, Epiroc, Cat, Outotec
Refractory (POX/roast)	Sulfide-locked gold	Very high	High	85-92%	Outotec, FLSmidth
Pressure oxidation (autoclave)	High-sulfide refractory	Highest	Highest	90%+	Specialty engineering

There is no universally best method. A company running multiple deposits across multiple methods (Newmont, Barrick, Agnico AEM) gets diversification but loses focus. A company running one or two assets on the same method (Wesdome WDO, Lundin Gold LUG) gets focus but takes single-asset risk.

The producer-stage margin is the headline number of the whole cycle: World Gold Council math implies industry margin of roughly $2,800 per ounce in 2026 against ~$1,500 in AISC, with 2025 industry-average AISC of $1,521/oz. At $4,700 gold and $1,500 AISC you are looking at $3,200/oz of gross cash margin. The producer is the only stage of the chain where this margin compresses or expands with the gold price — every other stage (equipment, reagents, refining fees, royalties as a percentage) is structurally insulated, which is why the upstream supplier oligopolies and the downstream royalty layer carry steadier margins than the miners they serve.

The parallel chain: royalty and streaming

Sitting alongside the physical chain is the financing chain. Franco-Nevada FNV, Wheaton Precious Metals, and Royal Gold provide upfront capital to miners in exchange for royalties on production or the right to buy gold at fixed prices. This is the highest-margin layer in the entire value chain — 70-80% gross margin, and for the best operator (FNV) the briefings cite 91% EBITDA margin and 82% FCF margin — because it strips out operational risk, capex risk, and permitting risk entirely while keeping near-1:1 operating leverage to the gold price. The model exists precisely because Stage 2 development capital is so expensive ($1.5-2.5B per new mine): the royalty/streaming companies are the capital-providers of last resort to a stage of the chain the miners themselves struggle to fund. Top 3 control ~80% of contract value. Diversification is the moat: no single FNV asset is more than ~13% of revenue, so a single mine going bad is a non-event — though the Cobre Panama shutdown (~15% of FNV revenue, shut by the Panamanian government in 2023) is the standing lesson that even the royalty layer inherits the political risk of its underlying mines.

Choke points and structural bottlenecks

The bottleneck that matters most for the bull case is permitting. A new mine in Nevada or Quebec takes 7-12 years from discovery to first pour; in Chile, Peru, or the US it can be 15+. This is the single most underappreciated bullish factor for incumbents: anyone holding permitted ore in the ground at ~$1,500 AISC has a moat that is essentially uncopyable for a decade. Permitting is the choke point that converts the geological supply constraint into durable incumbent pricing-power-by-scarcity (even though they have none on the gold price itself).

Reinforcing it is peak-gold geology: no major deposit (>2 Moz) found since 2022, average reserve grade for major producers down from ~1.5 g/t in 2000 to ~1.0 g/t today, deposits getting deeper, lower-grade, and harder to permit. Supply growth runs 1-2%/year and is forecast to peak around 2027-2028, then plateau or decline. The chain has two cycles operating at different timescales — the gold price cycle (5-7 years between peaks) and the capital cycle (a 7-10 year lag between price highs and new-mine completion) — and at present both are aligned bullishly: high price, no new supply for several years, record operating margins.

Other physical and structural choke points along the chain: the mill is the energy bottleneck (SAG/ball grinding is the heaviest opex line, and ventilation is the largest opex line in deep underground mines — the reason battery-electric LHDs are being adopted, to remove diesel and the ventilation it demands); mining tires as a chronic supply constraint; sodium cyanide as a duopoly-supplied critical consumable; tailings capacity, where the post-Brumadinho Global Industry Standard on Tailings Management (GISTM) forces more expensive dry-stack facilities; and the refining accreditation gate (LBMA Good Delivery) plus the geographic single-point-of-failure of Swiss refining concentration.

Why this chain doesn't get disrupted

The honest summary on technology: gold mining is a slowly improving heavy industry, productivity gains of 1-2%/year, and no new technology has changed the economics of gold mining since heap leaching in the 1980s. AI-driven exploration is a real productivity boost but does not create new ore bodies. Autonomous trucks, battery-electric underground equipment, in-situ leaching (not yet working for gold), and tailings reprocessing are all incremental. The investment case is not technology disruption — it is the price-margin spread on mostly the same operations, captured by whoever holds permitted, low-cost ore in a Tier 1 jurisdiction. Company-specific positioning within this chain lives on the ticker pages: AEM, AGI, EGO, WDO, LUG, FNV.

Players

The gold space splits into two structurally different businesses: the producers (the leveraged play on the gold price, 40-60% EBITDA margins at current spot but real operational, jurisdictional, and execution risk) and the royalty/streaming companies (70-80% gross margins, near-zero operational risk, the cleanest way to own gold-price exposure). The producer side is more fragmented than most assume — the top 10 producers account for only ~30% of global mine supply, and the single largest producer (Newmont) has roughly 6% share. There is no Saudi Aramco of gold. The picks-and-shovels layer is the opposite: Caterpillar, Komatsu, Hitachi, and Liebherr hold 80%+ of surface-mining trucks; Sandvik, Epiroc, and Cat hold ~75% of underground equipment.

The names below are organized by where they sit in the value chain and, for the Canadian no-Africa universe, by the head-to-head ranking from the screen.

The "no-Africa" Canadian universe — the core working set

This whole sector page exists because of a Doug filter: "Canadian-listed, no Africa, not the African one." Five names pass that filter, and they were screened head-to-head, deep-dived (the three finalists), and synthesized into a position-sizing recommendation. The head-to-head work lives in two documents worth referencing by name: the Gold No-Africa Screen (the five-name side-by-side data sheet, gold-no-africa-screen.md) and the Gold No-Africa Final Synthesis (the position-sizing recommendation, gold-final-synthesis.md). The no-Africa filter reflects Fraser Institute jurisdictional reality: West African mines have been seized by junta governments in Burkina Faso, Mali, and Niger over 2022-2025, and country risk that used to discount valuations 10-20% now discounts them 30-50%.

The screen's final ranking and the synthesis's position recommendation differ slightly — the screen ranked by setup, the synthesis re-ranked by post-deep-dive conviction:

Rank (screen)	Ticker	Setup	Conviction (post-DD)	Recommended size
1	AGI	Mid-tier GARP with brownfield catalyst	HIGH	3% over 3 tranches
2	EGO	Deep value with paired catalysts	Skipped (two simultaneous greenfield ramps)	—
3	AEM	Quality compounder, late in the cycle	Medium	3% on pullback only
4	LUG	Single-asset cash machine	Skipped (single-asset Ecuador)	—
5	WDO	High-beta basket trade	Medium (basket only)	1% over 3 tranches
—	FNV	Royalty ballast (added on top of producer picks)	High (royalty quality)	2% over 2 tranches

The recommended structure is 9% of portfolio split across one core grower (AGI), one royalty compounder (FNV), one opportunistic quality senior (AEM), and one beta basket (WDO) — skipping EGO and LUG, both addable later if execution unfolds well. Most-likely-Doug-pick was flagged as AEM (his style is quality-compounder, hold-for-years), but Pink's most-aligned pick if she trusts the numbers is AGI.

Cross-name comparison (the screen data)

Production & operations:

Metric	AEM	AGI	EGO	WDO	LUG
2025 production (oz)	3,450,000	545,400	488,268	185,576	498,315
2026 guidance (oz)	3,300,000-3,500,000	570,000-650,000	490,000-590,000	180,000-205,000	475,000-525,000
2027/2028 target	~flat through 2028	900,000+ by 2028	620,000-720,000 (2027)	continuous via expl.	similar to 2026
Growth profile	Flat to gentle	+65% by 2028	+40% by 2027	Stable	Stable
Number of mines	11	4	4 (5 with Foran)	2	1
% Canada	~75%	~70%	~38% (rises with Foran)	100%	0%

Cost structure (AISC):

Metric	AEM	AGI	EGO	WDO	LUG
2025 AISC ($/oz)	$1,339	$1,524	$1,664	$1,518	$1,015
2026 AISC guide ($/oz)	$1,400-1,550	TBD (rising)	$1,670-1,870	$1,525-1,700	$1,110-1,170
YoY trajectory	+12% creep	flat-to-up	+5% creep	+4% YoY	+10% creep
2028+ target	$1,400-1,500	<$1,025 mine-site	negative on Cu credits	depends on expl.	similar
Cash margin/oz at $4,700	$3,361	$3,176	$3,036	$3,182	$3,685

LUG is the cost king by over $300/oz; AGI is the only name that gets cheaper through 2028 as Island Gold Phase 3+ pulls mine-site AISC to ~$1,025/oz on the largest production block.

Reserves & mine life:

Metric	AEM	AGI	EGO	WDO	LUG
P+P reserves (Moz)	55.4	15.9	12.5	1.13	~6
Reserve life (yr)	~16	~25 (post P3+)	~18	~6	~12
2025 reserve growth	+2%	+32% (Magino)	+5%	growing	flat
Avg reserve grade (g/t)	~1.2	~1.4	~1.05	12.67	~9.5

WDO's six-year reserve life is the screen's single starkest number — it must drill its way out of every year. AGI's 25-year life post-Phase-3+ is exceptional. AEM grew reserves 2% in 2025 while mining 3.4 Moz, the rare case where exploration replaces depletion.

Balance sheet, valuation, risk (selected):

Metric	AEM	AGI	EGO	WDO	LUG
Market cap	$104.5B	$19.4B	$9.4B	C$4.05B	C$25B
Net cash/(debt)	+$2.5B	+$423M	-$406M (0.5x EBITDA)	+C$354M	+$630M
2025 FCF	~$4.4B	$352M (record)	~$700M	strong	$926M (~52% margin)
Dividend yield	0.86%	low	0.8% (initiated Jan 2026)	low	5-6% USD
P/E forward	15.3x	TBD	7.5x	TBD	TBD
EV/EBITDA	12.4x	TBD	TBD	6.1x	TBD
P/NAV	1.5-1.8x	1.2-1.4x	0.9-1.1x	1.1-1.3x	1.4-1.6x
Insider activity (12mo)	-$40M sells, $0 buys	Mixed	Mixed	TBD	Lundin family ~30%

EGO is cheapest on forward earnings (7.5x); WDO cheapest on EV/EBITDA (6.1x, ~10% forward FCF yield); AEM richest (1.5-1.8x P/NAV). The AEM insider selling — $40M sold over 12 months with zero buying — is flagged as the single most concerning data point in the screen, though the deep-dive could not confirm whether the sales were programmatic 10b5-1 plans (the open verification item).

Producer positioning, name by name

AGI — Alamos Gold (NYSE/TSX: AGI), the mid-tier growth story. Three North American mines (Young-Davidson and Island Gold in Ontario, Mulatos in Sonora), pure Canada/Mexico after selling its Türkiye projects for $470M in 2024, debt-free. The thesis is Island Gold, one of the highest-grade gold mines in the Western world (10+ g/t), and its Phase 3+ brownfield expansion — fully sanctioned, shaft sunk to 1,350m of 1,379m planned, 91% of growth capital already spent. The critical data point from the deep-dive: the project NPV works at $3,200 gold (Alamos's own base case), not the $4,500 the screen first modeled — NPV $8.16B, 53% IRR, project break-even sub-$1,200/oz. You're buying a brownfield-expansion compounder, not a leveraged gold bet. Production goes 545koz → 900koz+ by 2028 at sub-$1,025/oz mine-site AISC; 2026 is the FCF trough (capex peak $895M, FCF compressed to $50-100M) with FCF returning to ~$1.1B by 2028. Highest conviction in the universe; the one name where waiting costs optionality.

AEM — Agnico Eagle (NYSE/TSX: AEM), the quality compounder. Canada's largest pure-gold producer, #3 globally by output but the highest-quality portfolio of the top 5. Eleven mines, ~75% of production from Canada, the rest OECD; zero Africa, Russia, or non-Mexico LatAm exposure. Four cornerstone mines (Detour Lake, Canadian Malartic, Meadowbank, Meliadine) each 350-700koz/yr. Best management and capital allocation in the producer set — returned $1.4B to shareholders in 2025, raised the dividend 12.5% in Feb 2026, not chasing M&A. Q4 2025 was the best quarter in company history ($1.31B FCF in 90 days), and the balance sheet flipped to net cash (the synthesis puts it at +$2.5-2.67B). The bear case softened on deep-dive: the 12% AISC creep is "more than half" cyclical (royalty step-ups + CAD strength per CEO commentary), and reserves are struck at $1,600/oz — 35% of spot — so the 88.9 Moz combined M&I + Inferred is huge optionality. The catch is entry: the stock tripled in 24 months, insiders are selling, and the premium (1.5-1.8x P/NAV, 15.3x forward) compresses faster than discount-priced peers if gold pulls back. Own it — but on a -10% pullback, not at current levels.

WDO — Wesdome Gold Mines (TSX: WDO), the Canadian high-grade pure-play. Two underground mines, 100% Canada, 100% high-grade: Eagle River (Wawa, Ontario) and Kiena (Val-d'Or, Quebec, restarted 2022). Highest reserve grade in the listed universe at 12.67 g/t and the highest beta to the gold price — ~2.5x operating leverage. The investment angle is pure: if gold rips harder than expected, WDO outperforms by 2.5x; if it flatlines, WDO underperforms. The two structural weaknesses are a ~6-year reserve life and two-mine concentration (a geotechnical event at either site cuts production 40-50%). Q4 2025 AISC spiked to $1,750 on a Kiena hoist shutdown; 2026 guidance assumes normalization. The CEO change (Tyler Mitchelson, a heavier operator than Belleau) is a positive. The June 2026 NI 43-101 reserve update is the single biggest swing data point — the reserve-replacement story stands or falls there. A basket/beta sleeve position only, never a single-stock bet.

EGO — Eldorado Gold (NYSE: EGO / TSX: ELD), the Skouries story. Vancouver-headquartered mid-tier, 488,268 oz in 2025 across four mines in Canada (Lamaque, Quebec), Türkiye (Kisladag, Efemcukuru), and Greece (Olympias). The growth engine is Skouries, a long-delayed copper-gold project in northern Greece finally approved and slated for 2026 first production — adds ~140koz gold and ~67Mlbs copper at sub-$700/oz net of copper credits, driving the 2027 guidance jump to 620-720koz (+40%). Cheapest in the group on forward earnings (7.5x), the deep-value flag, with two simultaneous catalysts (Skouries Q4 2026 + McIlvenna Bay via the Foran deal mid-2026). The cost: ~40-50% of current production is Türkiye (currency + political risk), and the thesis hinges on two greenfield ramps landing at once — the reason the synthesis skipped EGO despite ranking it #2 on setup. Addable later on first-production confirmation.

LUG — Lundin Gold (TSX: LUG), the single-mine cash machine. One asset, Fruta del Norte in southeast Ecuador, ~498koz in 2025 at the best AISC in the group (~$1,015/oz). Highest-grade large mine after WDO (~9.5 g/t), built on time and budget, ~30% Lundin-family owned, structured as a dividend distribution machine yielding 5-6% in USD with ~52% FCF margin ($926M FCF, $664M returned). The whole thesis is cash flow today, not growth. The risk is total single-asset, single-country concentration — when the mine has a problem, the stock has a problem, and there is no second asset. Skipped in the synthesis for that reason; the consensus Hold reflects "price-in," not skepticism.

Royalty / streaming — the parallel chain

The royalty/streaming layer is a separate business: upfront capital to miners in exchange for a fixed share of production or the right to buy gold at fixed prices. Top 3 control ~80% of contract value. The economics are the most attractive in the whole chain — 70-80% gross margins, near-zero operational/capex/permitting risk.

FNV — Franco-Nevada (NYSE/TSX: FNV), the royalty king. The oldest, largest, and best-run royalty company in the world — the "Berkshire of gold." 91% EBITDA margin, 82% FCF margin, zero debt, $3.1B available capital, 19 consecutive annual dividend hikes, 430 assets (119 producing) with no single asset over ~13% of revenue. Compounds NAV per share at 10-12% annually for two decades. The investment angle is ballast: it strips out operational risk and compounds steadily through cycles — pair it with AGI as the growth name for a growth + quality + ballast structure. Two caveats: it trades at a premium (P/NAV 1.8-2.2x, ~28-43x P/E priced for continuation, so it compresses fastest if gold stalls), and Cobre Panama — a major royalty asset shut by the Panamanian government in 2023, ~15% of revenue — is treated as zero in 2026 guidance. If the ICC arbitration (hard date October 2026) ever resolves, that's free upside on the compounding franchise. The lesson: even royalty companies carry political risk through their underlying mines.

The other royalty/streaming names worth knowing, not deep-dived here:

Royalty/Streaming	Ticker	2025 Revenue	Market Cap	Notes
Franco-Nevada	FNV	$1.2B	$35-40B	Largest, most diversified, gold-base metals-energy
Wheaton Precious Metals	WPM	$1.0B	$30-35B	Streaming, gold + silver, growth-tilted
Royal Gold	RGLD	$700M	$12-15B	Smallest of big 3, US-listed, more concentrated
Triple Flag	TFPM	$200M	$3-4B	Mid-cap, growing
Sandstorm Gold	SAND	$200M	$2B	Junior-focused, more risk
Osisko Gold Royalties	OR	$250M	$3-4B	Tied to mid-tier producers

Rule of thumb: streamers compound mid-teens through cycles; miners compound high-teens-to-twenties in good cycles and zero in bad ones. A core-and-satellite approach (FNV core + AEM satellite) captures most of the upside with much less drawdown.

The seniors and the names that fail the filter

The full producer landscape, for context — the bolded "No" names are the no-Africa universe; the rest fail Doug's filter:

Company	Ticker	HQ	2025 Prod (Moz)	Reserves (Moz)	Africa exposure?	Tier
Newmont	NEM	US	6.0	135	Yes (Ahafo, Akyem)	Senior
Barrick Mining Corp	B / GOLD	Canada	3.9	89	Yes (Kibali, Loulo, Tongon)	Senior
Agnico Eagle	AEM	Canada	3.4	55.4	No	Senior
AngloGold Ashanti	AU	UK	2.6	30	Yes (Geita, Iduapriem, Obuasi)	Senior
Gold Fields	GFI	South Africa	2.2	47	Yes (South Deep, Damang, Tarkwa)	Senior
Kinross Gold	KGC	Canada	2.1	25	Yes (Tasiast = 25% of prod)	Senior
Northern Star	NST.AX	Australia	1.8	21	No (Australian-listed)	Senior
Endeavour Mining	EDV.TO	UK	1.1	17	Yes (pure West Africa)	Mid
Pan American Silver	PAAS	Canada	0.9 (gold)	15	No	Mid
Alamos Gold	AGI	Canada	0.55	15.9	No	Mid
Eldorado Gold	EGO	Canada	0.49	12.5	No	Mid
Lundin Gold	LUG.TO	Canada	0.50	6	No	Mid (single asset)
B2Gold	BTG	Canada	1.0	11	Yes (Mali, Namibia)	Mid
IAMGOLD	IAG	Canada	0.7	17	Yes (Burkina Faso, Mali)	Mid
Centerra Gold	CGAU	Canada	0.39	5	No (Türkiye + Canada)	Small
Wesdome Gold	WDO.TO	Canada	0.18	1.13	No (Canada-only)	Small
K92 Mining	KNT.TO	Canada	0.15	2.5	No (Papua New Guinea)	Small

The explicit avoids for the no-Africa filter: Barrick (Africa + management churn), Newmont (African + LatAm discount), Kinross (Tasiast is 25% of production), B2Gold (Mali Fekola), IAMGOLD (Burkina Faso). Northern Star is a sixth no-Africa name but Australian-listed; K92 carries PNG risk; Pan American is mostly silver; Centerra qualifies but Türkiye exposure has been volatile.

Emerging players and disruptors

The disruption pattern in gold is "permitted ore in a Tier 1 jurisdiction taken from a junior to mid-tier producer," not a Tesla moment — no new technology has changed the economics since heap leaching in the 1980s. Names to track:

• Skeena Resources (SKE) — Eskay Creek redevelopment in BC, high-grade gold-silver, first-pour 2027 target. The cleanest "next mid-tier from a permitted Canadian deposit" story.
• K92 Mining (KNT) — Papua New Guinea high-grade (Kainantu); exceptional asset, real country risk, Phase 3 expansion the catalyst.
• G Mining Ventures (GMIN) — Brazil, Tocantinzinho, the rare junior that builds on time.
• Calibre Mining (CXB) — Nicaragua + Nevada, growth profile, exits the Africa narrative.
• Snowline Gold (SGD) — Yukon discovery story (Rogue), no production yet.
• Junior streamers — Metalla, Vox Royalty, Empress Royalty: higher-risk, higher-beta versions of the FNV model.

The equipment and refining oligopolies (the picks-and-shovels)

Where the producer side is fragmented, the supplier side is a tight oligopoly — the secular, lower-beta way to play the cycle:

• Surface mining equipment: Caterpillar (CAT), Komatsu (6301), Hitachi Construction (6305), Liebherr (private) — the four-firm oligopoly, 70%+ of global surface fleet.
• Underground equipment: Sandvik (SAND.STO), Epiroc (EPI), Caterpillar — ~75% of underground trucks, ~88% of LHDs.
• Mill equipment (SAG/ball mills, crushers): Outotec/Metso (METSO), FLSmidth (FLS), Weir Group (WEIR) — the Big Three of mineral processing.
• Reagents: sodium cyanide is a quiet duopoly — Orica (ORI) and Cyanco/Draslovka (private). AIA Engineering (AIAENG) leads high-chrome grinding media.
• Drilling / assay: Major Drilling (MDI), Boart Longyear, ALS (ALQ, a quiet quality compounder), SGS, Bureau Veritas.
• Refining: ~60 LBMA-accredited refiners, Switzerland-dominated (refines ~two-thirds of newly-mined gold) — Valcambi, Argor-Heraeus, Metalor, PAMP (private Swiss), Rand Refinery, Royal Canadian Mint, Perth Mint, Tanaka (5901). Process margin is thin ($5-15/oz) but the LBMA Good Delivery List is a near-uncopyable moat.
• Exchanges / bullion banks: CME Group (CME) owns COMEX price discovery; LBMA market-makers (JP Morgan, HSBC, ICBC Standard, UBS) are not pure plays but JPM has the largest gold book.

ETF baskets

For basket exposure rather than single-name selection: GDX (VanEck Gold Miners, senior + mid-tier, the default), GDXJ (Junior Gold Miners, more leverage/volatility), SGDM (Sprott, quality-weighted to strong balance sheets), GOAU (US Global, heavy royalty/streamer weighting — the cleanest "premium quality" basket).

Monitor

A rolling log of dated developments, catalysts, and standing watch-items for the gold and precious-metals mining sector. Company-specific mechanics live on the ticker pages (AEM, AGI, EGO, WDO, LUG, FNV); this section tracks the dated, time-sensitive layer.

Dated catalyst calendar (set April 8, 2026 — next 90 days from then)

The whole stack of theses depends on gold staying above roughly $3,500. If it pulls back to $3,200 or lower, AGI still works on project economics but AEM and WDO price targets compress, and FNV compresses fastest because it was priced for continuation at ~43x P/E.

Catalyst	Date	Why it matters
FNV Investor Day, Toronto	April 8, 2026 (the briefing day itself)	New growth-pipeline detail; may move guidance. Deck expected on franco-nevada.com IR by end of day. Watch the reaction — if the stock holds or pops, initiate; if it pulls back, wait.
AEM Q1 2026 earnings	April 30, 2026	Cost-trajectory check, guidance update; insider blackout window opens after. The most actionable near-term event — if AEM beats again and insider selling pauses post-blackout, the buy case strengthens.
AGI Q1 2026 earnings	Early May 2026	Phase 3+ progress update, Magino mill ramp.
WDO Q1 2026 earnings	May 2026	Kiena normalization check; June reserve-update preview. If Q1 AISC is still elevated (vs the Q4 2025 spike to $1,750 on the Kiena hoist shutdown), the operating-leverage thesis cracks.
WDO June 2026 NI 43-101 reserve update	June 2026	The single biggest WDO catalyst — the reserve-replacement story stands or falls here. This is the decision point to size WDO up from 1% to 1.5% (if good) or trim to 0.5% / exit (if bad).
First Quantum / Cobre Panama ICC hearing	February 2026 (already occurred — watch the ruling)	FNV optionality unlock or stay-zero. Cobre Panama was shut by the Panamanian government in 2023, cost FNV ~15% of revenue, and is treated as zero in 2026 guidance.
FNV ICC arbitration	October 2026	The hard date for Cobre Panama clarity.
AGI Phase 3+ first ore (Island Gold)	Q4 2026	The AGI execution catalyst — confirms or breaks the thesis. Also the trigger for AGI tranche 3.
AGI Magino mill ramp	Q1 2028	Longer-term catalyst; the only thing that could derail Phase 3+. Track it.
EGO Skouries first production	2026 (Q4 target)	Greece's long-delayed copper-gold project; lifts EGO 2027 guidance from 488koz toward 620-720koz. Watch for confirmation before adding EGO.
EGO McIlvenna Bay / Foran deal	Mid-2026	The second of EGO's two simultaneous greenfield catalysts.
Skeena Resources / Eskay Creek first pour	2027 target	The cleanest "next mid-tier from a permitted Canadian deposit." Track until Q4 2026.

Standing watch-items (the recurring checklist)

These are the leading indicators the primer flags to read every cycle, not one-time events:

1. Gold ETF holdings (GLD, IAU). When North American ETF buying turns positive after years of outflows, that is the signal institutional money is rotating in. Miners lagged spot for the first time in two decades through 2024-25 — the 135% move in gold ($2,000 → $4,700) was matched by only ~100-120% in miners. The gap closing is the setup for the next 12-18 months.
2. Producer capex announcements. Watch for the moment management teams start saying "we're going to build new mines." That is the late-cycle sell signal.
3. Mining M&A activity. Small deals are healthy. A wave of large super-major combinations is the warning. Watch Indaba and Denver Gold Forum announcements specifically.
4. Real interest rates — 10-year TIPS yield, weekly. Below 1%, gold rallies; above 2.5%, gold struggles. A surprise tightening cycle would hit gold hard; the 2013 taper tantrum is the playbook to remember. This is the single biggest macro risk to the trade.
5. Central-bank buying continuity. Central banks bought 1,237 tonnes in 2025 (third consecutive year above 1,000t vs a pre-2022 average of 400-500t); 755 tonnes forecast for 2026; 76% of central banks plan to increase holdings over five years. China's PBoC bought every month for 15+ consecutive months as of early 2026. If China or Russia stops buying for two consecutive months, that is a major tell.
6. Recycling supply elasticity. 2025 saw record recycled supply (1,330 tonnes). At $5,000+ this could become a meaningful headwind to net new demand. Watch scrap flows.
7. AISC / cost inflation. Industry AISC went from ~$1,000 in 2020 to ~$1,521 in 2025. If cost creep continues, cash margins compress even at higher gold prices. World Gold Council math implies ~$2,800/oz industry margin in 2026 against ~$1,500 AISC — track whether that spread holds.
8. Next major discovery. No major (>2 Moz) deposit has been found since 2022. If a junior announces one, the whole exploration sector re-rates.
9. Jurisdictional risk repricing. Country risk used to discount valuations 10-20%; in 2026 it discounts 30-50%, driven by junta seizures of West African mines (Burkina Faso, Mali, Niger, 2022-2025). The post-2022 trend in every jurisdiction is "more for the state" — even Tier 1 Canada raised mining tax in some provinces in 2024-25.

Price-forecast backdrop (as of April 2026)

Gold traded around $4,700/oz in April 2026. The sober institutional desks, all revising up after two years of being wrong on the upside:

• Goldman Sachs: $5,400 by year end
• Wells Fargo: $6,100-6,300
• UBS: $6,200, with upside to $7,200

Base case across major bank desks: gold $5,000-6,500 by end of 2026, another 10-40% from then-current levels. The cycle read was mid-to-late innings of a multi-year bull market that began in late 2022.

Open questions outstanding (April 8, 2026)

These are the unresolved items the synthesis left open — each is a thing-to-watch until it closes:

1. What did Doug actually mean by the "Canadian, no-Africa" gold tip? Unblocks everything else; confirm directly before sizing. If he names a different name (Newmont, Barrick, Wheaton, Royal Gold) or an Australian name, re-screen with his actual filter.
2. Are AEM insider sales programmatic 10b5-1 plans? The deep-dive could not pull individual Form 4 filings to confirm. $40M sold over 12 months with zero buying is the most concerning data point in the screen. Pull from EDGAR, classify each transaction as 10b5-1 or discretionary, retabulate. If even half are plan-based, the AEM signal weakens materially and AEM moves up the conviction ranking.
3. Will FNV's Investor Day move guidance? Watch the slides from the April 8 event.
4. Where does gold go from here? The thesis stack depends on staying above ~$3,500.
5. What is happening at WDO's Kiena hoist? Q4 2025 AISC spiked to $1,750 on the shutdown; 2026 guidance assumes normalization. If Q1 2026 AISC is still elevated, the operating-leverage thesis cracks.
6. Pink's existing portfolio gold exposure — the 9% recommendation must net against any gold names already held.

What to read next (standing source list)

• World Gold Council "Gold Outlook 2026" (January 2026) — the industry's best free macro analysis.
• McKinsey "Can the gold industry return to the golden age?" — best supply-side analysis.
• Crescat Capital quarterly letters — Tavi Costa, loudest gold bull, charts usually right.
• Northern Miner / Mining.com daily — best industry news flow.
• Q4 2025 MD&A for AEM, AGI, EGO, WDO, LUG (all dropped February 2026).
• People to follow: Tavi Costa (Crescat), Lyn Alden, John Hathaway (Sprott), Pierre Lassonde (FNV founder), Adrian Day, Rick Rule, Bob Moriarty (321Gold), Doug Casey.

Sources

The gold-mining sector view is built on a chain of internal research documents plus a set of external references cited across them. The internal files are the consolidated source set; the external references are what the primer and screens leaned on for macro, supply-side, and cost data.

Consolidated source files (internal)

These are the vault documents this sector page synthesizes. Read them for the full reasoning, the per-mine tables, and the position-sizing logic.

• ~/Dropbox/Wafflebun/KB/wiki/gold-mine-supply-chain-primer.md — the industry primer (2026-04-07). End-to-end value chain, six-stage rock-to-bar flow, profit-pool map, players, trends, investment framework. The macro and structural backbone of the whole sector view.
• ~/Dropbox/Wafflebun/KB/wiki/gold-no-africa-screen.md — the five-name comparison screen (2026-04-07). Head-to-head of AEM, AGI, EGO, WDO, LUG on production, AISC, reserves, balance sheet, valuation, risk, and Doug-style fit.
• ~/Dropbox/Wafflebun/KB/wiki/gold-final-synthesis.md — the position-sizing synthesis (2026-04-08). Recommends 9% gold exposure split across AGI (3%), FNV (2%), AEM (3%, pullback-only), WDO (1%); the document Pink takes into the Doug conversation.
• The synthesis was itself built from the per-ticker /profile, /deep-dive, and /filings outputs: AEM/AEM.md, AGI/AGI.md, EGO/EGO.md, WDO/WDO.md, LUG/LUG.md, FNV/FNV.md, the three deep-dives (aem-deep-dive.md, agi-deep-dive.md, wdo-deep-dive.md), and the three filings notes (AEM-filings.md, AGI-filings.md, WDO-filings.md). Those live on the ticker pages — see the AEM, AGI, EGO, WDO, LUG, FNV wikilinks.

Briefing hubs (internal)

Six briefing stubs sit under KB/wiki/briefings/. Each is a short hub pointing at a long write-up rendered to the portal at pink.sakdiarpa.com/reports/. The full word counts live behind those URLs, not in the vault.

• briefings/2026-04-07-ego.md → EGO, ~8,417 words, portal reports/ego.html
• briefings/2026-04-08-aem.md → AEM, ~13,513 words, portal reports/aem.html
• briefings/2026-04-08-agi.md → AGI, ~19,462 words, portal reports/agi.html
• briefings/2026-04-08-fnv.md → FNV, ~7,637 words, portal reports/fnv.html
• briefings/2026-04-08-wdo.md → WDO, ~19,946 words, portal reports/wdo.html
• briefings/2026-04-08-gold.md → "GOLD" hub, ~4,808 words, portal reports/gold.html (the synthesis rendered as a portal briefing)

External references cited

The primer's own Sources block lists the following external authorities, with the data they anchored:

• World Gold Council — "Gold Outlook 2026", AISC Gold, Supply, Demand Trends 2025. The source for central-bank buying (1,237 tonnes in 2025; 755t forecast 2026), the ~$2,800/oz industry margin math, and 76% of central banks planning to raise holdings. https://www.gold.org/goldhub
• USGS — Gold Mineral Commodity Summaries 2025. https://pubs.usgs.gov/periodicals/mcs2025/mcs2025-gold.pdf
• LBMA — Good Delivery List and Responsible Sourcing Programme (the ~60 accredited refiners, 10-tonne/£15M/5-year accreditation bar). https://www.lbma.org.uk/good-delivery
• S&P Global — Mine Cost Outlook 2026 and AISC analysis (industry AISC $1,521/oz in 2025). https://www.spglobal.com/market-intelligence/.../mine-cost-outlook-2026
• McKinsey — "Can the gold industry return to the golden age?" (cited as the best supply-side analysis). https://www.mckinsey.com/industries/metals-and-mining/our-insights
• Mining.com — Top 10 Gold Companies 2025; also flagged as best daily industry news flow alongside Northern Miner. https://www.mining.com/ranked-top-10-gold-mining-companies-of-2025/
• Goldman Sachs / Wells Fargo / UBS — gold price forecasts (Goldman $5,400 year-end; Wells Fargo $6,100-6,300; UBS $6,200 with upside to $7,200). https://www.thestreet.com/investing/goldman-sachs-revamps-gold-price-target-for-the-rest-of-2026
• The Oregon Group — "Peak Gold: Is the world running out of gold?" https://theoregongroup.com/commodities/gold/peak-gold-is-the-world-running-out-of-gold/
• Globenewswire — mining-equipment market reports (Cat/Komatsu/Sandvik/Epiroc share data). https://www.globenewswire.com/news-release/2026/02/26/3245812/
• Company IR / results pages — Agnico Eagle (agnicoeagle.com), Alamos Gold (alamosgold.com), Eldorado Gold (eldoradogold.com), Wesdome (wesdome.com), Lundin Gold (lundingold.com), Kinross/Tasiast (kinross.com). Company Q4 2025 / FY 2025 earnings releases and investor decks were the per-name data source.
• Valuation / insider data feeds — Yahoo Finance / Morningstar for current multiples; SEDI / SEC EDGAR for insider transactions (the open item on confirming AEM's $40M of insider sells as 10b5-1 vs discretionary).

"What to read next" and people to follow

The primer also flagged a reading list and a people-to-follow list. These are recommended-but-not-yet-ingested sources:

• World Gold Council "Gold Outlook 2026" (Jan 2026) — the industry's best free macro analysis.
• Doug O'Laughlin on gold cycles — see source-doug-olaughlin (Fabricated Knowledge). The primer says "if he has written it up — search Fabricated Knowledge." Note: this is a different Doug from the "Doug" who supplied the no-Africa filter tip that triggered the whole gold workstream; that Doug is the tipster, not the Substack author.
• McKinsey "Can the gold industry return to the golden age?" — supply-side analysis.
• Crescat Capital quarterly letters / Tavi Costa — the loudest gold bull, macro charts.
• Northern Miner / Mining.com daily — industry news flow.
• People to follow per the primer: Tavi Costa (Crescat, macro gold bull), Lyn Alden (gold and monetary debasement), John Hathaway (Sprott), Pierre Lassonde (Franco-Nevada founder, the OG of royalty), Adrian Day (junior/exploration), Rick Rule, Bob Moriarty (321Gold), Doug Casey (contrarian).

None of these external people/publications except Doug O'Laughlin currently has a _sources/ page in the vault. If Crescat/Tavi Costa, Lyn Alden, or the World Gold Council become recurring inputs, they warrant their own source-*.md handles.

---

Consolidation queue (merged 2026-05-30 — section-scoped rebuild)

Industry-wide content folded in from these source files. They stay live pending Pink's archive confirm.

• [ ] gold-final-synthesis.md
• [ ] gold-mine-supply-chain-primer.md
• [ ] gold-no-africa-screen.md
• [ ] briefings/2026-04-07-ego.md
• [ ] briefings/2026-04-08-aem.md
• [ ] briefings/2026-04-08-agi.md
• [ ] briefings/2026-04-08-fnv.md
• [ ] briefings/2026-04-08-gold.md
• [ ] briefings/2026-04-08-wdo.md