THE SHORT ANSWER
China leads annual mine output; Australia and Russia lead the named reserve rows
The U.S. Geological Survey estimates that mines produced a rounded 3,300 metric tonnes of gold worldwide in 2025, up about 0.6% from the published 2024 total. China ranked first at 380 tonnes, followed by Russia at 310 and Australia at 280. The production ranking is not the reserve ranking: Australia and Russia together account for 25,000 of the 66,000 tonnes in the rounded world reserve estimate. Every 2025 production figure is estimated, and reserves are economic classifications rather than a count of all gold underground.
Rounded USGS estimate for 2025.
About 11.5% of the rounded world total.
Read the separate Australian reserve-basis footnote.
Annual source, not live mine telemetry.
INTERACTIVE WORLD MAP AND TABLE
Map two production years, then rank output, change, reserves or the static ratio
Choose 2024 or the estimated 2025 production view, switch to world share, annual change, reserves or the static reserve ratio, and inspect the 13 individually named countries by pointer, touch or keyboard. The complete source table, CSV and share link use the same versioned local data. Derived shares use published rounded world totals, so displayed components do not reconcile perfectly.
- 2024 production
- 377 t reported
- 2025 production
- 380 t estimate
- 2025 reserves
- 3,200 t
- Source status
- USGS estimate
The map places the 13 individually named USGS rows at approximate country centroids for orientation; it does not draw mine sites or allocate the “Other countries” aggregate. Use pointer, touch, Tab or arrow keys to inspect exact values.
| Country | 2024 production | 2025e production | Change | 2025 production share | 2025 reserves | Reserve / production |
|---|---|---|---|---|---|---|
| China | 377 t | 380 t estimated | +0.8% | 11.5% | 3,200 t | 8.4x |
| Russia2024 estimated | 310 t | 310 t estimated | 0.0% | 9.4% | 12,000 t | 38.7x |
| Australia | 284 t | 280 t estimated | -1.4% | 8.5% | 13,000 tSee Australia note | 46.4x |
| Canada2024 estimated | 200 t | 200 t estimated | 0.0% | 6.1% | 3,200 t | 16.0x |
| United States | 163 t | 160 t estimated | -1.8% | 4.8% | 3,000 t | 18.8x |
| Ghana | 149 t | 150 t estimated | +0.7% | 4.5% | 1,000 t | 6.7x |
| Mexico | 140 t | 140 t estimated | 0.0% | 4.2% | 1,400 t | 10.0x |
| Kazakhstan2024 estimated | 130 t | 130 t estimated | 0.0% | 3.9% | 2,300 t | 17.7x |
| Uzbekistan | 129 t | 130 t estimated | +0.8% | 3.9% | 2,200 t | 16.9x |
| Peru | 108 t | 110 t estimated | +1.9% | 3.3% | 2,200 t | 20.0x |
| Indonesia2024 estimated | 94 t | 90 t estimated | -4.3% | 2.7% | 3,600 t | 40.0x |
| South Africa | 90 t | 90 t estimated | 0.0% | 2.7% | 5,000 t | 55.6x |
| Brazil2024 estimated | 82 t | 80 t estimated | -2.4% | 2.4% | 2,500 t | 31.3x |
| Other countries | 1,020 t | 1,000 t | -2.0% | 30.3% | 11,000 t | Aggregate |
| World total (rounded) | 3,280 t | 3,300 t | +0.6% | 100% | 66,000 t | Not mine life |
Source: U.S. Geological Survey, Mineral Commodity Summaries 2026, gold chapter and companion data release, Version 1.3, May 2026. CSV verified against official MD5 36185ff3742087e1dd90c52fe634fe12. The 2025 production column is estimated; world totals and component rows are rounded. Map positions are approximate GoldObserve orientation coordinates, not USGS fields.
WHAT THE NUMBERS MEASURE
Mine output is not refinery production, recycling or official holdings
A country can mine ore, ship doré abroad for refining, import bullion, recycle jewelry and hold monetary gold at the same time. Those flows answer different questions. This page deliberately keeps the table on a gold-content, mine-production basis.
PRODUCTION CONCENTRATION
The five largest producers supply about two-fifths of published world output
The five leading 2025 rows are China, Russia, Australia, Canada and the United States. Their displayed estimates total 1,330 tonnes, or 40.3% of the rounded 3,300-tonne world total. The USGS narrative describes the same group as 41% of world production, a reasonable difference because the chapter table displays rounded country values while the narrative can reflect underlying estimates.
380 tonnes estimated; 11.5% of the rounded world total.
310 tonnes estimated; 9.4% of the rounded world total.
280 tonnes estimated; 8.5% of the rounded world total.
360 tonnes combined; about 10.9% of the rounded total.
The “Other countries” row remains large at 1,000 tonnes, or roughly 30.3% of the rounded world total. That aggregate prevents a false impression that the named leaders describe every producing jurisdiction, but it cannot support country-level claims inside the group.
YEAR-OVER-YEAR CHANGE
The published 2025 increase is small relative to annual output
Rounded world production rises from 3,280 tonnes in 2024 to 3,300 tonnes in 2025, a 20-tonne or approximately 0.6% increase. At the displayed precision, China rises by 3 tonnes, Ghana by 1, Peru by 2 and Uzbekistan by 1. The United States, Australia, Brazil and Indonesia decline modestly, while several named rows are unchanged.
These apparent movements should not be overinterpreted. Every 2025 row is estimated, five 2024 named-country rows are also marked estimated, and a one- or two-tonne change may be smaller than later revisions. The source table is appropriate for annual scale and rank, not for high-frequency supply signals.
The named countries plus the published Other countries aggregate sit 4 tonnes below the rounded 2024 world total and 50 tonnes below the rounded 2025 total. GoldObserve preserves the official rows and labels both residuals as rounding instead of allocating either difference to countries.
SUPPLY CHANGE RECONCILIATION
The world total rose, but the displayed country rows did not explain the increase
Preserving the residual is more truthful than forcing rounded components to equal the rounded world total.
2025 OUTPUT CONCENTRATION
Leaders matter, but “Other countries” remains almost one-third
Top 3: 29.4% · Top 5: 40.3% · Top 10: 60.3%. Shares use displayed estimates and the rounded world total.
What the chart supports: the named rows show four increases, four decreases and five unchanged displayed estimates. It does not support assigning the 20-tonne world increase to a country because the rounded residual changed by 46 tonnes.
Interpretation boundary. The residual is not measured output and must not be allocated to a country. Mine production excludes recycled supply, refinery throughput and above-ground stock changes. Annual estimates and concentration shares are not price signals or forecasts.
RESERVES VS. PRODUCTION
Large annual output does not imply the largest reserve base
China has the largest 2025 production estimate but 3,200 tonnes of published reserves. Australia produces less in the year yet has a 13,000-tonne published reserve figure, while Russia has 12,000 tonnes. South Africa's 5,000 tonnes of reserves also sit far above its 90-tonne annual production estimate. The mismatch reflects geology, project maturity, economics, classification and reporting basis—not an automatic statement about future production.
3,200 tonnes of reserves divided by 380 tonnes of estimated output equals 8.4x.
13,000 divided by 280 equals 46.4x, before applying the separate JORC-equivalent note.
12,000 divided by 310 equals 38.7x.
5,000 divided by 90 equals 55.6x.
Australia requires special care. USGS publishes 13,000 tonnes in the reserves column and separately states that JORC-compliant or equivalent reserves were 4,500 tonnes. A comparison that silently replaces one basis for some countries but not others can reverse rankings, so the explorer retains the headline USGS column and exposes the footnote.
THE STATIC RATIO
Reserve divided by production is a comparison tool, not a depletion clock
The result has a years-like unit only because tonnes are divided by tonnes per year. GoldObserve labels it a ratio, not remaining mine life.
The calculation assumes production never changes and ignores new discoveries, drilling, mine plans, permitting, closures, ore grades, metallurgical recovery, costs, exchange rates and gold prices. It also assumes reserve definitions are comparable. None of those assumptions is realistic over a multi-decade horizon.
Use the ratio to identify where published reserve scale is high or low relative to one year's output. For an actual mine-life study, use asset-level technical reports, proven and probable reserve categories, planned throughput, recovery rates, sustaining capital, closure schedules and jurisdiction-specific disclosure standards.
WHY RESERVES CHANGE
Economic reserves can rise even while mines extract gold
Exploration can convert material into measured or indicated resources and later into reserves.
Discovery is not immediate production.Higher prices or lower costs can make additional material economic; the reverse can remove it.
Reserve growth can be an economic reclassification.Processing technology, dilution, recovery, throughput and design affect extractable tonnes.
Ore tonnes are not the same as contained gold.Government, company and reporting-code updates can change the published basis.
Compare dates and footnotes before ranking.The 2026 USGS gold chapter explicitly says reserves for Australia, Brazil, China, Peru and Other countries were revised using company and government reports. A change between annual tables therefore cannot be attributed to mining alone.
PRODUCTION, RESERVES AND REVISIONS
Three evidence layers answer three different questions
Annual mine production records newly extracted gold content for a period. Mineral reserves classify material considered economically extractable under a stated basis. Later source editions can revise either history or classification. GoldObserve keeps those layers separate before calculating rankings or static ratios.
MARKET INTERPRETATION
Annual mine supply matters, but it is not a mechanical gold-price signal
Gold differs from a commodity that is mostly consumed after production. A very large above-ground stock can return through recycling, investment sales, official transactions or inventory movement. Annual mine production adds new metal, but price formation also reflects jewelry and technology demand, bars and coins, exchange-traded products, central-bank behavior, derivatives, real interest rates, currencies and risk preferences.
A small global production increase can coexist with a rising price when demand or desired inventories increase faster. Production can fall while price also falls if investment demand weakens or holders sell existing gold. The country table is best used to understand supply geography, geological-economic capacity and revision risk—not to convert tonnes directly into a price target.
01Use annual production to describe new mine supply, not total tradable gold.
02Keep production estimates separate from actual releases and later revisions.
03Compare reserve definitions before drawing depletion conclusions.
04Add recycling and demand evidence before making a market-balance claim.
05Use macro and positioning data separately; do not force one-variable causality.
06State what observation would invalidate the supply narrative.
Calculation detailsEstimate flags, aggregates and rounding
METHODOLOGY AND REPRODUCIBILITY
Every displayed value traces to the USGS chapter and companion CSV
Select one commodity chapterGold only; units are metric tonnes of gold content.
Preserve source flagsEvery 2025 production row is estimated; selected 2024 rows are also estimated.
Separate aggregatesOther countries and World total are not ranked as countries.
Calculate without re-rounding inputsChange, share and ratio use the values exactly as published in the table.
Expose residualsRounded component rows are not forced to equal rounded world totals.
Version the sourceUSGS MCS 2026 version 1.3 and data release updated May 27, 2026.
The downloaded GoldObserve CSV includes source values, estimate flags, derived changes and shares, the Australian reserve note and aggregate rows. The official source CSV was verified against MD5 36185ff3742087e1dd90c52fe634fe12. Derived percentages are GoldObserve calculations and may differ slightly from narrative percentages calculated from unrounded USGS estimates.
Official sources and licenceUSGS MCS 2026 and companion data
PRIMARY SOURCES AND LICENSE
Official USGS publication, data release and definitions
- USGS Mineral Commodity Summaries 2026, version 1.3, for publication history, scope, citation and the statement that MCS is the earliest comprehensive source of 2025 world mineral-production data.
- USGS Gold chapter for the 2024 and estimated 2025 mine-production table, 2025 reserves, narrative concentration statement and footnotes.
- USGS MCS 2026 companion data release for the machine-readable commodity records used to build the explorer.
- USGS data catalog record for update date, public access, provenance and the public-domain label.
- USGS Appendix C for mineral resource and reserve classifications and source conventions.
The underlying USGS data release is identified as public domain. GoldObserve adds calculations, interface and explanatory analysis, preserves attribution and does not imply USGS endorsement. Later USGS revisions should replace, not silently blend with, this source version.
RELATED RESEARCH
Connect mine supply to ownership, prices and market drivers
FAQ
Gold production and reserve questions
Which country produces the most gold?
China ranks first in the USGS 2025 estimate at 380 metric tonnes of mine production, followed by Russia at 310 tonnes and Australia at 280 tonnes. All 2025 production values in the table are estimates.
How much gold was mined worldwide in 2025?
USGS estimates rounded world mine production at 3,300 metric tonnes in 2025, up from a rounded 3,280 tonnes in 2024. That is an increase of about 0.6 percent based on the published rounded totals.
Which country has the largest gold reserves?
Among the individually named rows, Australia has the largest published reserve figure at 13,000 tonnes, followed by Russia at 12,000 tonnes. USGS separately notes that Australia's JORC-compliant or equivalent reserves were 4,500 tonnes, so the basis matters.
Does gold mine production include recycled gold?
No. Mine production is newly extracted gold content. Recycled gold, secondary refinery production, imports, central-bank transactions and movement of above-ground stocks are separate supply or ownership categories.
Are 2025 gold production numbers final?
No. The USGS table marks every 2025 country production value as estimated. Several 2024 country values are also estimated, and later editions can revise historical values as company and government reports improve.
What do gold reserves mean?
Mineral reserves are the economically extractable part of a measured or indicated mineral resource under the assumptions and reporting basis used by the source. They are not all gold in the ground, and they can change with prices, costs, technology, drilling, regulation and classification.
Is the reserve-to-production ratio the number of years until a country runs out of gold?
No. Dividing reserves by one year's estimated production is only a static ratio. It assumes unchanged production and no discoveries, reserve revisions, project additions, closures, cost changes or economic reclassification, so it is not a mine-life forecast or depletion date.
Why do the country rows not add exactly to the world total?
USGS labels the world totals as rounded. The named country rows and the Other countries aggregate are also presented at limited precision, so their displayed sum can differ from the rounded world figure. GoldObserve preserves that residual instead of forcing the rows to reconcile.
Does higher mine production make the gold price fall?
Not automatically. Mine supply changes slowly and is only one part of the market. Recycling, investment flows, jewelry and technology demand, central-bank activity, real interest rates, currencies, inventories and expectations can outweigh a small annual production change.