Are Gold Miners Leveraged to Gold? Barrick's Quarter Says 1.7x, Not 3x

The question of whether gold miners are leveraged to gold has two confident answers in circulation, and they contradict each other. One of them was published this morning by a company that digs the metal out of the ground.

The familiar answer is two to three times. Gold climbs 30%, a miner's margin climbs 60% to 90%. The reasoning behind it is clean. It costs roughly the same to lift an ounce out of the ground whether that ounce sells for $2,000 or $4,000, so the extra price falls through to profit nearly untouched. Barrick reported its second quarter before the open today, and the filing carries both halves of the test. Its realized gold price came in at $4,417 an ounce against $3,295 a year earlier, a rise of 34%. Its margin per ounce went from $1,614 to $2,551, a rise of 58%.

That is 1.71 times.

Leverage is real. It is also a long way from three, and the distance between those two numbers is the cost line.

Ava, the Kodex guide who reads structure and pressure before she reads price, settles the argument with one subtraction. She does not start with the miner. She starts with what the miner is selling you.

What a gold miner actually sells you

You own the miner instead of the metal, and this morning the headline reads net earnings of $1.22 billion, up 50% year on year, with adjusted earnings per share up 74%. It looks like the thesis working. Ava's first question is not whether the number is good. It is which part of it repeats the next time gold moves, and which part was simply gold moving this time.

To separate those, she strips the release down to two numbers: what the company received for an ounce, and what the ounce cost to produce and keep producing. The gap between them is the margin per ounce.

Hold the metal and your return is the price change. Nothing sits in between. Hold the miner and your return runs through a subtraction, because the company has to buy the ounce out of the ground before it can sell it to anyone.

Industry calls that second number all-in sustaining cost, written AISC. It bundles the cash cost of mining with the sustaining capital, site administration and royalties needed to hold production steady. It leaves out growth projects, so it is not the full cost of the business, but it is the closest published figure to what an ounce really costs to keep making.

So the leverage question is a question about one line of arithmetic: realized price minus AISC equals margin per ounce. Everything else in the argument is decoration on that subtraction.

The case that gold miners are leveraged 2x to 3x

Two to three times is not folklore. It follows from an assumption that is reasonable on its face, and if the assumption held, the number would be right.

Assume AISC stands still. Barrick's cost a year earlier works out to about $1,681 an ounce. Barrick reports only the change, not the old number, so that figure comes from dividing this year's $1,866 by 1.11. Freeze the cost there, apply this year's $4,417 realized price, and the margin becomes $2,736. Against last year's $1,614, that is a rise of 69.5% on a 34% move in the metal.

That is 2.04 times, from the same filing, on the same quarter.

That model is arithmetically sound, and it is built on something true. Digging costs are dominated by things with no connection to the gold price: diesel, wages, haul trucks, the depth of the pit. A price rise arrives as revenue on top of a cost base set by geology and by contracts signed years ago. VanEck's commentary on gold miner margins starts from the same place before adding its caveats.

The assumption is the whole argument. Barrick's quarter is a direct test of it.

Barrick indexed its own cost forecast to the gold price

Ava stops on one sentence in the guidance section and reads it twice. Nothing in the results table matters as much.

Barrick's second quarter release puts full-year AISC at $1,760 to $1,950 an ounce, and then states that the guidance is "based on a gold price assumption of $4,500 per ounce."

A company cannot forecast what an ounce will cost to produce without first assuming what an ounce will sell for. If costs were independent of price, that sentence would not need to exist.

That same release names the mechanism outright. Gold cost of sales came in at $1,993 an ounce against $1,654 a year earlier, "primarily due to lower grades processed at Carlin, Cortez, and North Mara; higher fuel costs across the operations; and higher royalties associated with the stronger realized gold price." Two of the three drivers in that sentence are the cost line responding to price: royalties rise because the metal is worth more, and grades fall because a higher price makes poorer rock worth digging.

The company is not hiding the feedback loop. It is publishing it inside the cost guidance.

Why does part of every price rise get spent buying it back?

Ava splits the drivers into two families, because they bite at different speeds and lumping them together makes the effect look like bad luck.

The first family is revenue-linked. These rise automatically because the metal is worth more, with nothing changing underground. Sliding-scale royalties are the clearest case: VanEck notes that some governments set royalty rates that climb as the gold price climbs, so the state's share widens on its own. Employee profit-sharing behaves the same way. More profit generated, more profit shared.

The second family is price-induced, and it is slower and stranger. Here the higher price changes what the company chooses to do, and the choice raises the cost.

Cutoff grade is the one worth sitting with, because it runs against instinct.

Every deposit holds rock of varying richness. A miner sets a cutoff: rock above this grade is ore, rock below it is waste. That line is not geological. It is economic, and it moves with the price. Take a tonne of rock carrying half a gram of gold, about one sixty-fourth of an ounce. At $2,000 an ounce that tonne holds roughly $32 of metal, which does not cover hauling and crushing it, so it is waste. At $4,400 the same tonne holds about $71, and it pays its way. Nothing about the rock changed.

So the miner processes it. Total ounces produced go up, which is what a shareholder wanted. Cost per ounce also goes up, because each ounce now takes more rock, more crushing, more diesel, more hours.

The reserve grew and the margin per ounce thinned, and both happened for the same reason.

Barrick's release names exactly this at three of its largest operations. VanEck adds three further price-induced drivers: sector inflation as mining companies compete for the same rigs and crews, producer-currency appreciation that lifts US dollar costs when local currencies strengthen, and capital spending accelerating because the free cash flow arrived to fund it.

None of this is mismanagement. It is what a rational company does when the price of its product rises.

Mining difficulty runs the same feedback loop

A crypto-first reader has already met this mechanism under a different name.

Bitcoin's price rises. Mining gets more profitable. Rigs that were uneconomic at the old price get switched on, hashrate climbs, difficulty adjusts upward, and the cost of producing one bitcoin rises to meet the new price. The protocol does it on a fixed schedule. Gold does it through royalty formulas, cutoff grades and diesel invoices. The ledger is different and the feedback is identical.

That same arithmetic decides whether a lone machine ever wins a block, which is the subject of solo Bitcoin mining odds. In both markets a price rise summons the marginal producer, and the marginal producer is expensive by definition.

Are gold miners leveraged to gold on your own numbers?

"You do not have to accept anyone's ratio, including mine," Ava says. The test runs on four lines and one quarterly release.

You pull the realized price for both quarters and the cost per ounce for both, subtract to get the margin per ounce in each, then compare the percentage change in margin against the percentage change in realized price. That ratio is the leverage the company actually delivered.

Barrick's second quarter, worked through:

A 34% move in the metal produced a 58% move in the margin, so the amplification holds. It also fell 0.33x short of what a frozen cost line would have produced, and that shortfall is roughly 11 percentage points of margin growth absorbed on the way through.

Three limits belong in the open, because a test is only as honest as its inputs.

That $1,681 is derived rather than published. Barrick reports "up 11%", which is rounded, so the base is approximate. Running the sensitivity settles it: at 10.5% the answer is 1.73x, at 11.5% it is 1.68x. The rounding band cannot stretch far enough to rescue a 2x claim.

There is also a second route that needs no derivation at all. Cost of sales is published for both quarters, $1,993 against $1,654. Margin measured that way went from $1,641 to $2,424, a rise of 47.7%, which lands at 1.40x. Two cost definitions, two different answers, both under two.

And this is one company in one quarter. It is evidence that 2x to 3x is not a constant. It is not evidence that 1.7x is one either.

Where the 1.7x stops being true

Margin per ounce is not the share price, and collapsing the two is the fastest way to misuse everything above.

Between the margin and the stock sit production volume, the multiple the market assigns to those earnings, jurisdiction risk, debt, and whether the company sold its output forward. A producer that hedged at $3,000 captured none of the move to $4,417. Its margin arithmetic says one thing and its shareholders lived through another.

Royalty and streaming companies inverse the problem entirely. They buy a slice of future production at a fixed price and carry almost no operating cost line, so their leverage to gold is cleaner than a producer's for structural reasons rather than managerial ones.

Barrick is not an outlier in the direction, though. The World Gold Council put industry AISC at a record $1,706 an ounce in the fourth quarter of 2025, up 20% year on year. It attributed the rise to higher operating costs, more sustaining capital and less gold sold, "as did royalty payments tied to the gold price" while that price ran up 55%. The feedback shows up at industry scale, in the council's own words.

What transfers to the next earnings print is the reading habit. When a headline reports a mining company's earnings jumping, the number that decides whether the jump repeats is the cost line. It is the same discipline that showed gold overtaking Treasuries in central bank reserves was price rather than buying. It is why a stock drops after good earnings on the strength of the forecast rather than the quarter. A published number is where the arithmetic starts. The one-cent rule that sets ETF spreads hides in the same place, one line below the figure everyone reads.

Ava closes the release and leaves the two percentages sitting on the screen: 34 and 58. The gap between them is not a rounding error or a weak quarter. It is the standing price of owning the company that digs instead of the metal it digs.

The arithmetic is yours now. Test it where both positions sit on one screen: hold gold and a tokenized stock on the same $5,000 paper account in the simulator, ride a single macro headline, and watch how differently the two of them arrive. Kodex lists no mining companies, so the lesson is not the pair, it is the habit of asking what stands between a price and a profit.

Start free →