How much is left, and how long will it last?

At the pace the world mined in 2025, the reserves counted for 2025 would last about 18 years for zinc, 20 for gold and 21 for lead, and more than a century for 9 of the 29 listed below. None of those is a date when anything runs out: for 14 of the 18 minerals we can check, the reserve figure is higher than it was three years earlier, although about three years of mining happened in between.

The table answers four questions for each material: how much is mined each year, how much is left in reserves, how many years that is at the current pace, and how much has already been taken out. Reserves are the part of a deposit that can be mined at a profit with the technology and prices of the year counted, as USGS and the Energy Institute count them. They are a working inventory, not the amount in the ground, which is why the number of years can lengthen even while mining continues.

Mined each year, left and years at the latest year's pace

Click a heading to sort. "Years left" is at the 2025 pace; "years if output grows" assumes production continues to change by the average yearly rate of 2013–15 to 2023–25 and is shown only where that rate is positive. "Mined so far" is world production summed from the year shown to 2025; "known stock" is mined so far plus reserves.

Mined each year and reserves for minerals: USGS Mineral Commodity Summaries 2026 (the same figures as each commodity page). Oil, gas and coal: Energy Institute Statistical Review of World Energy 2026, whose reserves were last updated for the end of 2020; the years shown for them take away what was produced in 2021–2025 and divide by 2025 output. History: USGS Data Series 140 and the Mineral Commodity Summaries 2022–2025. ≥ marks reserves that USGS prints as "more than" a figure (an open-ended world total), so the years are a minimum. Quantities are in each row's own unit (tonnes, barrels or cubic metres), so the quantity columns sort by their value in tonnes: coal as given, oil using the Energy Institute's own tonnes figure, and natural gas, which has no mass figure in the data, sorts to the end.

How much of what was ever found has been mined?

Each bar splits the known stock, everything mined in the years on record plus the latest reserves, into what has already been taken out (dark) and what reserves still hold (light). It is a floor on how much was found worth mining, not the geological total, which nobody knows, and it only counts the years shown after "since". A high bar means most of what has been identified has been mined; it does not mean the rest is about to be exhausted, because reserves grow when new deposits are found or become profitable.

Reserves are not a fixed stock

If reserves were a tank, each year's mining would lower the level by that amount. The record says otherwise. Comparing USGS's reserve figure in the 2023 Mineral Commodity Summaries (a 2022 figure) with the 2026 edition, and subtracting what was mined in 2023, 2024 and 2025: reserves are larger in the latest edition than in 2022 for 14 of 18 minerals (zinc, gold, tin, lead, silver, fluorspar, cobalt, copper, iron ore, tungsten, molybdenum, manganese, lithium, phosphate rock); and for 4 they fell by more than was mined (bauxite, graphite, vanadium, rare earths), which mining alone cannot explain. Over a longer span the Energy Institute counts proved oil reserves rising from 683 billion barrels in 1980 to 1,732 billion in 2020, a period in which about 1,093 billion barrels were produced, and gas from 70.9 to 188 trillion cubic metres.

"Years then" is reserves at the earlier date divided by that year's production; "years, latest" is the figure from the first table (the latest reserves divided by 2025 production; for oil and gas, end-2020 reserves less what was produced since). Minerals: USGS Mineral Commodity Summaries 2023 and 2026, same basis in both tables (a series is left out when the two tables are not comparable or give only a lower bound). Oil and gas: Energy Institute, 1980 to 2020. Reserve figures move for several reasons, new discoveries, new methods, changing prices and re-estimation among them; the table shows that they move, not why.

What these numbers can and cannot tell you

• "Years left" is a ratio, not a forecast. It divides the latest reserves by one year's mining. It would be a date only if reserves never changed and mining never changed, and neither is true.
• Reserves are narrower than resources. USGS separates reserves, which can be mined at a profit under the conditions of the year counted, from resources, which include deposits that are known but not yet economic or are only inferred. Resources are larger; USGS gives resource estimates in the text of its commodity chapters rather than in its data tables, so they are not counted here.
• Growing demand shortens the time. The "if output keeps growing" column applies each material's recent yearly growth to the same reserves, using the standard formula years = ln(1 + R × g / P) / ln(1 + g), where R is reserves, P is current production and g is the growth rate. It is arithmetic on a past trend, not a prediction of where mining is heading.
• Mining is not consuming. Metals can be recycled; coal, oil and gas are burned. The table counts only what comes out of the ground.
• Known stock is a floor. It adds reserves to the production of the years on record, so it leaves out anything mined before the first year shown and anything never counted as a reserve. The share "already mined" would be smaller with either included.

Why some rows have no "mined so far"

A cumulative total is only worth showing if the history can be joined to the latest figures without a gap or a change of basis. Where it could not, the row keeps its reserve life and shows a dash. Nothing was estimated to fill the gap.

Antimony

USGS's world antimony series breaks in 2019-2020 (162,000 t, then 111,000 t on a revised basis), so the history cannot be joined.

Chromium

The historical series is chromium content; current production and reserves are marketable chromite ore, so the two cannot be added.

Barite

USGS's barite world totals exclude the United States until MCS 2025 and include it after, so the series changes basis.

Titanium

The historical ilmenite series ends in 2020 and the later MCS tables do not report world ilmenite on the same basis.

Iron ore

The historical series counts usable ore; reserves are compared with production as contained iron, so the two cannot be added.

Bauxite

USGS's historical statistics have no world bauxite series.

Coal

The Energy Institute's coal production series starts in 1981; coal had been mined for centuries before that, so a total since then says little about how much was there.

Platinum-group metals

The historical series counts all platinum-group metals; current production here is platinum plus palladium only, so the two cannot be added.

How it is built

• Mined each year and reserves, minerals: USGS Mineral Commodity Summaries 2026 (production and reserves). Each commodity page shows the same two figures; this page reads them from there, so the two cannot disagree. Where the two are on different bases the row uses a matching pair: iron ore compares contained-iron reserves with contained-iron output, which is why its reserves here differ from the crude-ore figure on its commodity page. Platinum-group metals compare USGS's reserve figure for all platinum-group metals with platinum and palladium output only, so their years of supply are somewhat high.
• Mined so far: world production from Kelly, T.D., and Matos, G.R., comps., Historical statistics for mineral and material commodities in the United States: U.S. Geological Survey Data Series 140 (USGS), which runs back to 1900 for many minerals, then the settled figure for each later year from the matching Mineral Commodity Summaries (2022–2025) until the 2026 edition. The sum starts at the first year of the longest unbroken run that ends in 2025, shown in the table.
• Oil, gas and coal: Energy Institute, Statistical Review of World Energy 2026. Its reserves were last updated for the end of 2020, so the page subtracts what was produced since. Its own reserves-to-production ratios for 2020 were checked against ours for every fuel and agree within 1%.
• Data: /data/depletion.json holds every row, with the reserve date, the years of history and the reason where a figure is left out. Related reports: where reserves sit against where they are mined, critical minerals.

Try your own assumptions in the reserve scenario calculator.

Figures read on 6 October 2026; none of this is investment advice.