Gold 22 Percent Below Its Record, Newmont at an All-Time High: The Arithmetic of a Gold Mine — and Who Really Owns the Leverage

Gold is down roughly 22 percent from the record it set on January 29, 2026. Newmont, the largest gold producer in the world, closed at an all-time high on August 25, 2026. Both statements are true, and anyone who reads them as a contradiction has never taken the arithmetic of a gold mine all the way to the end. That arithmetic fits on a napkin, and it is treacherous enough that the industry used it to destroy roughly $80 billion between 2011 and 2015. This analysis shows why an ounce of gold at $4,300 is worth more to Newmont than at any point before 2026, where roughly a quarter of every price increase quietly disappears on the way to the bottom line, why Barrick generated less free cash flow at $4,400 gold than it did a year earlier at $3,300 — and what a price-to-earnings ratio of 15 assumes about a company whose most important number is set by someone else.

Buying a gold miner is, in theory, buying leverage to the gold price. In practice that leverage ran the wrong way for twelve years: in mid-2023 gold stood where it had stood in September 2011, and the GDX miners index stood roughly 55 percent lower. Since 2024 the leverage has been working again, and spectacularly so — GDX returned about 110 percent in the twelve months to the end of March 2026, against 52 percent for the metal. Whether that is the rule or the exception is not decided by the gold price. It is decided by three numbers that appear in every quarterly report and that almost nobody reads together.

The napkin: price minus cost, and the rest is leverage

A gold mine sells a product whose price it does not set, at a cost it only partly controls. Profit per ounce is the gap between the realized gold price and the so-called all-in sustaining cost, or AISC — what it takes to keep current production running at today’s level. Newmont has guided to roughly 5.26 million ounces in 2026 at an AISC of $1,680 per ounce. Everything else is arithmetic.

At $3,300 gold, the level of the second quarter of 2025, a cost of $1,680 leaves $1,620 per ounce. At $4,300, where the metal traded in mid-September 2026, it leaves $2,620. Gold rose 30 percent; the margin per ounce rose 62 percent. At the January record of about $5,400 on the London fix, the margin would have been $3,720 — up 130 percent from the year-earlier level. That is the leverage, and it grows the closer the gold price sits to cost: a mine with $1,680 of cost at $2,000 gold doubles its margin when gold rises 16 percent.

Gold price ($/oz)	Change in gold	Margin per ounce at $1,680 AISC	Change in margin	Newmont, 5.26 million oz ($ billion)
3,300 (Q2 2025)	—	1,620	—	8.5
3,600 (bear case)	+9%	1,920	+19%	10.1
4,300 (September 2026)	+30%	2,620	+62%	13.8
5,000 (bull case)	+52%	3,320	+105%	17.5
5,400 (January 2026 record)	+64%	3,720	+130%	19.6

The table explains Newmont’s all-time high at a gold price a fifth below its own record: the margin per ounce at $4,300 is still 62 percent higher than a year ago, and the market prices cash flow, not the metal. In the second quarter of 2026 Newmont generated $2.2 billion of free cash flow at a realized price of $4,414 — a record for a second quarter — and in the first quarter, at $4,900, it generated $3.1 billion. The first half of 2026 therefore produced $5.3 billion of free cash flow, more than the company earned in some entire years of the previous decade.

The table also shows the flip side that gets lost in the euphoria. The leverage runs both ways: if gold falls from $4,300 to $3,600, a 16 percent decline, the margin shrinks 27 percent. And the table assumes something that is never true in practice — that costs stand still while the price moves.

What AISC measures — and what it deliberately leaves out

All-in sustaining cost is a child of the last bust. The World Gold Council introduced the metric in 2013 after years in which the industry had advertised so-called cash costs that excluded sustaining capital, corporate overhead and exploration. A mine with $600 cash costs could lose money at $1,200 gold, and the investor found out only in the annual report. AISC was meant to end that: it includes production cost, royalties to the host government, general and administrative expense, exploration on existing properties and the capital required to hold production flat.

It leaves out three things, and all three cost money. First, growth capital: what a new mine or an expansion of an old one costs is not in AISC. Second, interest and taxes. Third — and this is the subtlest gap — by-products. Newmont reports second-quarter AISC of $1,621 per ounce. That figure is struck after crediting the revenue from the copper, silver, lead and zinc that come out of the ground alongside the gold. Allocate costs across all metals instead, the so-called co-product method, and Newmont’s AISC is $1,938 per ounce — $317 or 20 percent higher. Both numbers appear in the same release; the lower one appears in the headline.

The difference is not academic. A fifth of the leverage an investor thinks they are buying in Newmont is leverage to the copper price. And the first quarter of 2026 showed how much the by-product method swings: Newmont reported AISC of just $1,029 per ounce, because high copper and silver prices and a favorable production mix inflated the credits. One quarter later the figure was $1,621, a 58 percent jump, without the mines having become fundamentally more expensive. Anyone who took the first-quarter number as the cost base overestimated the leverage by a third.

Three producers, one gold price, three completely different outcomes

Nothing exposes the limits of the napkin better than the second quarter of 2026, in which the three large North American producers sold their gold at practically the same price — between $4,414 and $4,483 per ounce — and reported three businesses that could hardly have less in common.

Q2 2026	Newmont	Barrick	Agnico Eagle
Gold production (thousand oz)	1,293	796	856
Realized gold price ($/oz)	4,414	4,417	4,483
AISC ($/oz)	1,621 (co-product: 1,938)	1,866	1,459
AISC, year-earlier quarter	1,375	1,684	1,281
Cost increase year over year	+18%	+11%	+14%
Margin per ounce (price minus AISC)	2,793	2,551	3,024
Operating cash flow ($ billion)	2.92	1.70	2.14
Free cash flow ($ billion)	2.21	0.14 (attributable)	1.34
Net cash ($ billion)	3.4	1.25	3.27
Returned to shareholders in the quarter ($ billion)	1.9	1.5	0.63

Agnico Eagle has the lowest cost at $1,459, mines almost exclusively in Canada, Finland, Australia and Mexico, and turned 856,000 ounces into $1.34 billion of free cash flow — $1,560 per ounce produced. Newmont, with a bigger portfolio and higher costs, generated $1,710 of free cash flow per ounce, but unlike Agnico it sold billions of dollars of mines last year and has cut its share count by more than nine percent since February 2024. Barrick, at practically the same gold price, turned 796,000 ounces into $141 million of attributable free cash flow, or $177 per ounce — less than the $212 million of the year-earlier quarter, when gold fetched $3,295.

Barrick’s gold price rose 34 percent; its free cash flow fell by a third. That is the number that proves the leverage is a simplification. Barrick invested $1.19 billion in the quarter, $654 million of it in projects that do not yet produce an ounce: the Reko Diq copper-gold deposit in Pakistan, the Lumwana Super Pit in Zambia, the Pueblo Viejo expansion in the Dominican Republic and the Fourmile project in Nevada. At the same time, grades at Carlin and Cortez fell, diesel got more expensive and royalties rose with the gold price: cost of sales per ounce was $1,993, 20 percent above the prior year. Barrick earns more from operations than ever — $1.70 billion of operating cash flow, up 28 percent — and spends it the same second on the next decade.

Whether that is wise will be decided between 2028 and 2030, when Reko Diq and Lumwana are supposed to deliver. For today’s investor it means this: buying Barrick is not buying leverage to the gold price, it is buying leverage to the gold price four years from now, discounted for the risk that Pakistan, Zambia and Mali honor their contracts. The market values that at roughly $71 billion — about half of Newmont for roughly 60 percent of the production.

Where a quarter of every price increase goes

Newmont’s AISC rose from $1,375 to $1,621 year over year, by $246 or 18 percent. Its realized gold price rose $1,094 over the same period. Of every dollar the metal gained, 22 cents flowed into cost and 78 cents into margin. At Agnico the split is $178 of $1,195, or 15 cents; at Barrick $182 of $1,122, or 16 cents. The industry-wide average that Metals Focus compiled for the first quarter of 2026 stands at $1,785 per ounce — 16 percent above the prior year and the 28th consecutive quarter of rising costs.

That costs rise with price is neither an accident nor sloppiness. It has four causes, and they act in a particular order.

The first is the state. Royalties — revenue-based levies paid to the host country — are typically defined as a percentage of the gold price. When the price rises the levy rises automatically, with no legislature having to vote. According to Metals Focus, royalties rose 85 percent year over year in the first quarter of 2026; their share of AISC doubled from six percent in early 2021 to twelve percent. And legislatures vote anyway: Ghana, Africa’s largest producer, has introduced a sliding scale that takes up to twelve percent of revenue at prices above $4,500. It is the tobacco logic of gold mining: the state is a silent partner, and its share grows with the price.

The second cause is energy. A gold mine is a diesel business — Agnico Eagle puts diesel at roughly ten percent of operating cost, and all three producers cite fuel as a 2026 cost driver. The third is labor and consumables, from explosives to sodium cyanide, in an industry that competes worldwide for the same engineers and geologists.

The fourth cause is the most interesting because it is a decision, not an imposition: ore grade. A mine is not a warehouse from which gold is withdrawn but a deposit with rich zones and poor ones. At $2,000 gold, rock carrying 0.8 grams per tonne is not worth processing; at $4,300 it is. So producers lower the so-called cut-off grade, move more tonnes with less gold in them — and cost per ounce rises even though cost per tonne does not. Barrick’s reference to lower grades at Carlin and Cortez is partly exactly that: a response to the high price, not depletion. It is the right business decision, because it extends mine life and raises absolute profit. But it means unit costs structurally chase the gold price, on the way up and on the way down. When the price falls, producers raise the cut-off again, mine the rich zones first and AISC declines. The leverage is therefore smaller in both directions than the napkin promises — and in a downturn the industry protects itself by eating its own future.

The lost decade: what the leverage does when nobody is watching it

To understand why gold miners were one of the worst asset classes in the world for twelve years despite this arithmetic, visit 2011. Gold reached about $1,900 in September of that year, and the industry did what industries with suddenly fat margins always do: it invested as though the price would rise forever. Barrick bought the copper producer Equinox for $7.3 billion in 2011 and pushed ahead with Pascua-Lama on the Chile-Argentina border, whose cost estimate rose from three billion dollars to more than eight. Kinross had paid $7.1 billion for Red Back Mining in 2010. Newmont expanded.

Then gold fell 28 percent in 2013 and the arithmetic ran in reverse. Barrick reported a 2013 net loss of $10.4 billion, including $11.5 billion of impairments, six billion on Pascua-Lama alone, whose construction was halted. The company cut the gold price it uses to calculate reserves from $1,500 to $1,100 — and erased roughly a quarter of its reserves, because the ore was no longer economic at that price. Between 2011 and 2015 the five largest gold producers wrote off roughly $80 billion combined, essentially on acquisitions made at peak prices and projects whose costs had run out of control.

What followed was a decade of penance. The industry slashed capital spending from 2015 to 2021 even as global output grew about 20 percent; the cost of discovering a new ounce more than doubled over the decade, according to Sprott; the gap between discovery and first cash flow lengthened from roughly six years to more than ten. And because producers repaired their balance sheets with equity issuance, they diluted shareholders at exactly the moment their shares were cheapest. The result: gold was back at roughly $1,950 in mid-2023, its 2011 level. GDX sat about 55 percent below its high of that year.

The lesson of that decade is the most important number in this analysis, though it appears in no table: a gold mine’s leverage acts on the margin per ounce, not on the shareholder. Between the margin and the shareholder stand management and its decisions on acquisitions, projects and share count. From 2011 to 2015 those decisions consumed the leverage entirely.

What is different in 2026 — and what is not

The strongest argument for gold miners in 2026 is not the gold price but what the producers are doing with the money. All three majors carry net cash: Newmont $3.4 billion, Agnico $3.3 billion against just $197 million of debt, Barrick $1.25 billion. Newmont has repurchased more than 100 million shares since February 2024, nine percent of the count, with $4.3 billion left of a $6 billion program; in the second quarter of 2026 alone $1.7 billion went to buybacks. Agnico bought back 2.24 million shares at an average $179 and pays a quarterly dividend of 45 cents. Barrick spent $1.21 billion on buybacks and more than tripled shareholder returns year over year.

This is a different industry from 2012. Producers book reserves conservatively: Newmont raised its reserve price for 2025 from $1,700 to $2,000 — up 18 percent, at a spot price more than double that. Only ore that would still be economic if gold halved sits in the books. And the acquisitions that did happen — Newmont bought Newcrest for $16.8 billion in 2023 — were financed afterward by selling the weaker mines, not by issuing shares.

What is not different: the temptation. Barrick’s $654 million of project capital in a single quarter is exactly the 2011 pattern, only with better balance sheets. And the planned separation of Barrick’s North American business — Nevada Gold Mines, Pueblo Viejo, Fourmile and the interests acquired from Newmont, with an IPO targeted by the end of 2026 — is a bet that the market pays a premium for mines in safe countries that the remaining business in Mali, Pakistan and Zambia does not receive. It may work; it is also an admission that a third of the portfolio is a valuation burden.

The reserve as a clock: why every mine is a business with an expiry date

The second reason the margin arithmetic is not the whole truth lies in a number Newmont reported in February 2026: 118.2 million ounces of gold reserves, down from 134.1 million a year earlier. A decline of 12 percent in a year in which the gold price rose roughly 70 percent.

Newmont reserves 2025 (million oz)	Change
Balance, end of 2024	134.1
Mines sold	−8.6
Depletion (mined)	−7.2
Reclassified to resources (mainly Yanacocha Sulfides)	−5.6
Higher cost assumptions	−3.1
Higher reserve price ($1,700 → $2,000)	+6.6
Additions from exploration and conversion	+2.0
Balance, end of 2025	118.2

Two lines in that table deserve attention. Newmont mined 7.2 million ounces and found or converted 2.0 million. On a sustained basis the company replaces less than a third of what it produces. And even in a year of record gold prices, higher cost assumptions struck 3.1 million ounces from reserves — ore that was economic at $2,000 gold and 2025 costs and no longer is at 2026 costs. A reserve is not a quantity; it is a price ratio.

At 5.26 million ounces a year, 118 million ounces last 22 years on paper. That sounds like a lot and is not, because the 22 years are unevenly distributed: the richest zones are mined first, grades decline, and every ounce in the second half costs more than every ounce in the first. That is why Sprott pointed out in 2023 that analysts value gold miners with a model that has no terminal value — unlike for any other company, the firm is assumed to cease to exist after the last reserve ounce. It is also why a price-to-earnings ratio of 15 at Newmont is not comparable with a P/E of 15 at a consumer-staples company: one has to dig its earnings up again every year, the other does not.

What the stock price assumes: 15 times earnings at a gold price nobody guarantees

Newmont trades in mid-September 2026 at about $124, a market capitalization of roughly $131 billion and a price-to-earnings ratio of about 15.5 on trailing twelve-month earnings. First-half free cash flow of $5.3 billion annualizes to roughly eight percent of the market cap — at realized gold prices between $4,400 and $4,900. At the current $4,300 the sustainable yield is closer to seven percent.

Agnico Eagle is valued at roughly $100 billion; on annualized free cash flow of about $5.3 billion that is a yield of a little over five percent. Barrick comes to roughly $71 billion, and its free-cash-flow yield is, as shown, not a meaningful number right now because the company is redirecting cash into projects. The three valuations tell the same story in three keys: the market pays the highest premium for the lowest costs in the safest countries and applies the deepest discount to the most growth in the riskiest places.

Mid-September 2026	Newmont	Barrick	Agnico Eagle
Market capitalization ($ billion, rounded)	131	71	ca. 100
2026 production (million oz, guidance midpoint)	5.26	3.08	3.40
2026 AISC guidance ($/oz)	1,680	1,760–1,950	1,400–1,550
Market value per ounce of annual production ($ thousand)	25	23	29
Q2 free cash flow, annualized, as % of market value	6.7	0.8	5.3
Main producing countries	U.S., Australia, Canada, Ghana, Peru, PNG	U.S., Dominican Rep., Mali, Tanzania, Zambia, Pakistan	Canada, Finland, Australia, Mexico

What sits inside a P/E of 15? If Newmont produces the expected 5.26 million ounces at $1,680 in 2026, it earns roughly $13.8 billion of AISC margin at $4,300 gold. Subtract $1.4 billion of growth capital and roughly a third in taxes and what is left is broadly eight to nine billion dollars of free cash flow, consistent with the first-half run rate. At a $131 billion market cap the company is therefore valued at about 15 times the cash flow it generates at $4,300 gold — and at about 25 times what it would earn at $3,300, the price 15 months ago. The stock price assumes $4,300 stays. That is not an absurd assumption, but it is an assumption, and it is not within Newmont’s power.

Bull and bear case

The World Gold Council’s outlook for the second half of 2026 has a base case around $4,100, plus or minus five percent; an upside case of $4,500 to $5,000; and a downside of ten to 15 percent. The scenarios translate to Newmont — with the crucial addition that costs do not stand still in either direction.

Newmont 2027, illustrative	Bear case	Base case	Bull case
Gold price ($/oz)	3,600	4,300	5,000
AISC ($/oz)	1,700 (higher cut-off, lower royalties)	1,800 (+7%)	1,900 (more royalties, lower cut-off)
Margin per ounce	1,900	2,500	3,100
AISC margin on 5.3 million oz ($ billion)	10.1	13.3	16.4
Free cash flow, rough (after growth capital and taxes, $ billion)	ca. 5.5	ca. 8	ca. 10
Yield on $131 billion market value	4.2%	6.1%	7.6%
Share price if the market demands a 6.5% cash-flow yield ($)	ca. 80	ca. 115	ca. 145

The table is an illustration, not a forecast, but it makes the asymmetry visible. A gold price of $3,600 — 16 percent below today and still ten percent above last year’s average — would cut Newmont’s free cash flow by roughly a third and, at an unchanged valuation, the share price by roughly 35 percent. The bull case at $5,000 yields a gain of roughly 17 percent. The distribution is not symmetric because the stock already prices the base case, and because rising prices bring rising royalties and falling cut-off grades that dampen the leverage on the way up.

There is a second counterargument that weighs more than the gold price: the return of temptation. The industry is earning more money in 2026 than ever before, and history knows no commodity cycle in which record profits did not end in acquisitions at record prices. Newmont’s Newcrest purchase in 2023, Barrick’s Reko Diq build and the $1.95 billion top-up Newmont is paying to expand the Nevada joint venture are each defensible on their own. Taken together they are the 2011 pattern in slow motion. Anyone holding gold miners has to monitor capital allocation every quarter like a financial metric — because it is one.

Gold or gold miner? What the difference means for U.S. investors

The napkin leads to a question every gold investor should answer before buying a mining share: do I want the gold price, or do I want leverage to the gold price that is diluted by cost inflation, management decisions and political risk? Those who give the first answer belong in bullion or a physically backed fund such as GLD or IAU. Those who give the second buy miners — and should know that the U.S. tax code tilts the decision in a way most investors do not expect.

For U.S. taxpayers, physical gold and the ETFs that hold it are treated as collectibles. Long-term gains on collectibles are taxed at a maximum federal rate of 28 percent, not the 15 or 20 percent that applies to long-term gains on stocks — and that includes GLD and IAU, because the trusts hold bullion and the shareholder is treated as owning a pro-rata share of it. Gold mining shares and mining ETFs such as GDX, by contrast, are ordinary equities: long-term gains are taxed at 0, 15 or 20 percent depending on income, plus the 3.8 percent net investment income tax where it applies, and qualified dividends get the same preferential rates. On a hypothetical 50 percent gain held longer than a year, a high-bracket investor keeps 36 points of it in bullion and 40 to 42.5 in miners. The miner has a built-in tax edge in the United States — the mirror image of Germany and Austria, where physical gold is tax-free after twelve months and mining shares are not.

Within an IRA or 401(k) the distinction disappears, which is one reason the collectibles rule matters less than it looks for long-term holders. It also disappears for the one asset class that has neither cost inflation nor a mine life: royalty and streaming companies such as Franco-Nevada, which BMInsider examined in a separate deep dive in August. They are the attempt to buy leverage to the gold price without the cost side of the napkin — at a correspondingly higher multiple.

Those who want miners have three routes. Individual shares of the three majors trade on the NYSE; Agnico and Barrick are Canadian, so a 15 percent Canadian withholding tax applies to dividends in taxable accounts, generally creditable against U.S. tax and waived in IRAs under the treaty. A miners ETF such as GDX bundles roughly 50 producers and removes the single-name risks from Mali to Pakistan, but not the industry risks. And the royalty companies are the third route, for investors who have read this far and decided the cost side is the part they would rather not own.

The three conditions under which the leverage belongs to the shareholder

The arithmetic of a gold mine is simple, which is exactly why it is so often read wrong. Price minus cost is the margin per ounce; margin times production is gross cash flow; and at a gold price 30 percent above last year, that margin is 60 percent higher than a year ago even though gold has lost a fifth from its record. That explains Newmont’s all-time high, and it explains why GDX rose 110 percent in twelve months while gold rose 52.

But the arithmetic holds only under three conditions, and all three lie outside the gold price. First, costs must rise more slowly than the price — in 2026, 15 to 22 cents of every additional dollar goes to royalties, diesel, wages and lower grades, and that share grows with every sliding-scale royalty a host country enacts. Second, management must pass the cash flow on to shareholders rather than sink it into projects at the ends of the earth; Newmont and Agnico are doing so, Barrick is not, and the record of 2011 to 2015 says discipline erodes with the length of the boom. Third, the company must be able to replace its reserves without the replacement ounces costing more than the ones it mined — and Newmont’s ledger of 7.2 million ounces depleted against 2.0 million added says that is the hardest part.

Investors who check all three conditions in every quarterly report own leverage to the gold price. Investors who watch only the gold price own leverage that belongs to somebody else — the host government, the diesel supplier or the next acquisition target. The difference between the two was roughly 55 percent from 2011 to 2023. It will be again.