Two routes that meet at one doré bar
Nearly every route to pure gold or silver converges on the same object: an impure, mixed gold–silver bar called doré. Getting there is the hard part; splitting the two metals apart afterwards is chemistry that has barely changed in a century.
Gold and silver sit near the bottom of the reactivity series, which is exactly why they were the first metals humans ever used. They turn up as the pure metal, they don't rust, and they don't need a furnace to free them from oxygen the way iron does. That same nobility is the problem at industrial scale. You cannot burn or reduce these metals out of their ore, because they are barely bonded to anything. They are also present in vanishingly small amounts, often just a few grams of metal in a tonne of rock.
So the making of gold and silver is really a story of concentration: move mountains of rock, coax the metal into a liquid it will dissolve in, pull it back out as a solid, and only then purify it. The journey runs in four moves: free the metal from the rock, pull it out of solution, pool it into a doré bar, and finally part and refine the two metals to 99.99% ("four nines") purity. A second, quieter route supplies most of the world's silver: it rides out of the ground as a passenger inside copper, lead and zinc ores, and is caught later at the smelter.
Tap a stage to explore
Ore — and other metals' ore
The single most important fact about silver: most of it isn't mined for its own sake. By industry estimates roughly 70–80% of new silver is a by-product of mining copper, lead and zinc. So the amount the world gets is set less by how much silver we want than by how much of those base metals we happen to be digging.
Start with grade — how much metal is actually in the rock. Gold ore that is worth mining can carry just a few grams per tonne; silver ores are richer but still typically hold under 2% silver (Silver Institute). Those numbers are the whole reason the industry looks the way it does: to get a wedding ring's worth of gold you move something closer to a tonne of rock. The entire process is built around handling enormous volumes cheaply.
Gold is mostly a primary product, dug for itself from lode (hard-rock) and placer (river-gravel) deposits. Even here, though, about 7% of U.S. output comes as a by-product of copper mining (USGS, 2026). Silver is the opposite. In the United States only four mines produce silver as their main product; another 31 operations recover it as a by- or co-product of base- and precious-metal mining (USGS, 2026). Worldwide, silver comes chiefly from lead–zinc, copper and gold ores, in that order, and polymetallic deposits hold more than two-thirds of known silver (USGS, 2026).
The consequence is worth pausing on, because it drives the whole supply story later: if silver mostly arrives as a passenger, then mine supply can't simply be turned up when silver gets expensive. It moves with the copper, lead and zinc market instead.
Gold ore
Lode and placer deposits, dug for the gold itself. Grades are measured in grams of metal per tonne of rock — the reason everything downstream is built for volume.
Silver-bearing polymetallic ore
Galena (lead), sphalerite (zinc) and copper sulphides that carry silver as a passenger. It is recovered when those base metals are smelted and refined — not by mining silver on its own.
Copper anode slimes
The sludge left at the bottom of the cell after copper is electro-refined (purified with electric current) — a major primary source of both gold and silver. Follow it back and you are at the copper page.
Two roads, one destination
Because the metal is too noble to smelt out, the trick is chemical, not thermal: dissolve the gold or silver into a liquid that will hold it, leave the worthless rock behind, then bring the metal back out as a solid.
Crush → leach → recover → doré
Smelt → catch the slimes → doré
However the metal was won, the routes meet at doré: a bar that is mostly gold and silver but still mixed together and not yet pure. Two jobs remain — separate the two metals from each other, and lift each to market purity.
Whichever route delivered it, doré still has to be parted and refined. Gold is chlorinated (the Miller process) then electroplated (Wohlwill). Silver is electro-refined in a nitrate bath. And where the two must be split chemically, nitric acid dissolves the silver and leaves the gold. The end products are “four nines” bars — 99.99% pure — and, for the most demanding electronics, 99.999%.
Two very different metals
Here the two metals split. Silver is an industrial metal that also stores value; gold is a store of value that is only incidentally industrial. Nearly all the gold ever mined still exists, sitting in jewellery and vaults; a large share of silver is used up and thrown away.
Gold's uses are dominated by holding wealth, not doing work. Of global gold demand (setting aside exchange-traded funds), jewellery takes about 40%, physical bars about 24%, and central banks and other institutions about 21%; coins add another 7%. Electrical and electronics — the only genuinely industrial slice — is just 7% (USGS via World Gold Council, 2026). That is the striking fact to leave with: for all its reputation as a high-tech metal, gold is barely used to make anything.
Silver is the mirror image, and on the same global footing the contrast is sharp. Of world silver demand in 2024, industrial fabrication took a record 680.5 million ounces — close to 59% of the total (Silver Institute, World Silver Survey 2025). Jewellery and silverware together came to roughly a fifth, net physical investment in coins and bars to about a sixth, and old-style photography to just 2%. Silver conducts electricity and heat better than any other metal and reflects light better too, which is why the industrial slice is so large. Solar cells alone consumed about 198 million ounces, close to a third of all industrial silver, with electronics the single biggest use of all.
Put the two side by side, both measured globally, and the shapes invert. Gold is something the world holds: more than nine-tenths of it goes into jewellery, bars, coins and central-bank reserves, and only about a fourteenth is used to make things. Silver is something the world uses up: close to three-fifths is transformed in industry. And where gold has a huge official-sector buyer — central banks and institutions take about a fifth of gold demand — silver has essentially none. Its price leans far more on factories than on vaults.
Gold — what the world does with itglobal · 2025
Silver — what the world does with itglobal · 2024
Both bars are measured globally, so the two are directly comparable. Gold: USGS via the World Gold Council (2025 data, excluding ETFs). Silver: Silver Institute, World Silver Survey 2025 (2024 data). Non-industrial silver shares are rounded. In the United States the silver mix tilts even more industrial — USGS puts electronics, other industry, solar and brazing at well over half of domestic demand — but the global figures are the like-for-like comparison with gold.
The scale, and the gap
The world mines far more silver than gold by weight — about eight tonnes of silver for every one of gold — yet gold is worth vastly more per ounce, which is why a year's gold output is a far larger prize than a year's silver.
The two metals come from different places. Mexico leads silver; China leads gold — and the scales are worlds apart.
Top silver producers tonnes · 2025
Top gold producers tonnes · 2025
Mine production, 2025 (USGS). Each chart is scaled to its own leader, so note the axes differ by roughly seventeen-fold: the top silver producer mines around 6,300 tonnes, the top gold producer about 380. The top five gold producers together account for about 41% of world output.
Two independent authorities line up, which is a useful sanity check on numbers that come from different methods. USGS and the Silver Institute's World Silver Survey both put 2024 world silver demand near 36,000 tonnes (about 1.16 billion ounces) and mine output near 26,000 tonnes.
A note on reserves. USGS puts known silver reserves near 610,000 tonnes and gold reserves near 66,000 tonnes. Divide silver reserves by annual mine production (≈ 610,000 ÷ 26,000) and you get a reserve life of roughly 20–25 years — a figure the refining literature also cites (emew, 2025). That is a ratio, not a forecast: reserves are what is economic to mine today, and they tend to grow as prices rise and exploration continues. So "20 years left" should be read as a measure of tightness, not a countdown.
Asked to do more, as supply tightens
Silver is being asked to do more — by solar panels and electronics above all — at exactly the moment its supply is hardest to increase, because most of it depends on how much copper, lead and zinc the world decides to mine.
For years now the silver market has been in deficit. In 2024 the world used about 149 million ounces (~4,600 tonnes) more silver than mining and recycling together supplied — the fourth consecutive annual shortfall. The 2021–2024 deficits add up to roughly ten months of global mine supply (Silver Institute). The gap is drawn down from above-ground stocks, and it was cited as one reason for the price climbing through 2025 (USGS, 2026). The demand side is easy to explain: a solar panel needs silver paste to carry its current, and solar has gone from about a tenth of industrial silver demand in 2014 to nearly a third by 2024 (Silver Institute). Electrical and electronic uses, photovoltaics among them, are the metal's single largest end-use.
The bind is that silver supply cannot easily answer. Because most silver is a by-product, opening a "silver mine" to meet silver demand rarely makes sense — the metal mostly comes up whether or not anyone wants more of it, tied to the economics of copper, lead and zinc. In late 2025 the U.S. government added silver to its list of critical minerals for the first time, formal recognition that a metal long thought of as merely precious is now strategically industrial (USGS, 2026).
There are two pressure valves, and neither is simple. The first is using less: solar manufacturers have been steadily "thrifting" silver — cutting the amount in each cell — which eases demand per panel but risks losing performance, and the sheer growth in panels can outrun the savings. The second is recycling: it recovered about 194 million ounces in 2024, a twelve-year high, but that is still only around a sixth of demand (Silver Institute), because so much silver ends up thinly spread through electronics and solar panels that are awkward to take apart. Gold faces the gentler version of the same question — it is valuable enough that little is ever truly lost, but pulling microscopic amounts back out of circuit boards is fiddly work.
The gold flecks are the silver — it comes up with copper, lead and zinc, not on its own. So when silver runs short you cannot simply mine more: supply is set by the base-metal market, not by silver's price. That is the bind the two valves below have to work around.
Use less — thrifting
Solar makers keep cutting the silver paste in each cell. It eases demand per panel — but the sheer growth in panels keeps outrunning the savings.
Recycle more
Recovery hit a twelve-year high in 2024 — but silver is spread so thinly through electronics and panels that most is still awkward to reclaim.
Underneath all of it sits the physical cost of the primary route: moving and grinding tonnes of rock for a few grams of metal takes real energy, and dissolving that metal means managing cyanide and the tailings left behind — the environmental price of chasing something this dilute. How much of tomorrow's silver comes from yesterday's solar panels, rather than from fresh rock, is one of the open questions of the energy transition.