3,300 Tonnes or 3,672? Why USGS and Metals Focus Disagree

Two of the most respected sources in gold reporting published materially

different 2025 supply figures. The USGS's Mineral Commodity Summaries

put global mine production at 3,300 tonnes; Metals Focus put it at 3,672

tonnes. The 370-tonne gap is not an error — it is a window into how the

gold market is measured, and why the answer matters for investors.¹²

The Two Headline Numbers

The US Geological Survey's February 2026 Mineral Commodity Summaries

estimated worldwide gold mine production at 3,300 tonnes in 2025, up

marginally from 3,280 tonnes in 2024.¹ Metals Focus, the specialist

consultancy whose open coverage appears across the Silver Institute,

Investing News Network and other outlets, reported 2025 mine production

at a record 3,672 tonnes.²

The difference is roughly 11.3 percent of the USGS figure — substantial

enough to matter for anyone sizing the market. Both numbers were

published in early 2026, both cover the same calendar year, and both

come from methodologies that are widely trusted. The disagreement is

structural, not episodic.

Where the Gap Comes From

The first driver is scope. The USGS reports what it defines as mine

production — gold content extracted from primary gold mines and

recovered as a by-product of base-metal operations. The number is

compiled from government statistical agencies, company reports and trade

data, and it reflects a conservative preference for reported rather than

estimated figures.

Metals Focus' 2025 total includes by-product gold that the USGS may

count only partially, and in some jurisdictions it incorporates volumes

from artisanal and small-scale mining (ASM) that national statistics do

not fully capture. For countries where informal mining is material —

parts of West Africa, the Tapajós region in Brazil, the eastern Congo

basin — the gap between statistical gold and actual gold mined can be

several dozen tonnes on its own.

A second driver is timing. The USGS's MCS is published in February with

a six-to-eight-week lag on calendar data; Metals Focus can incorporate

later revisions and primary-channel intelligence. When the two are read

together, the later number tends to be the larger one simply because

more information has been rolled in.

A third driver is recycling and treatment of refinery output. Some

refining statistics blur the line between primary mined gold and

recycled gold passing through the same facility. Both methodologies are

careful about the distinction, but the accounting choices around

refinery flows do affect the margins.

A fourth driver is the treatment of by-product gold from large copper

operations. The Sossego and Salobo mines in Pará, for example, primarily

produce copper but recover significant gold. Depending on whether a

national agency attributes that gold to a gold line in its trade

statistics or to a separate copper-concentrate line, it can move in and

out of the USGS tally in ways that Metals Focus' channel checks are

designed to catch more completely.

USGS Methodology in Brief

The USGS is a federal scientific agency. Its Mineral Commodity Summaries

is the earliest comprehensive source of production data for more than

ninety minerals worldwide, and its methodology is deliberately

conservative. The agency's gold chapter is prepared by a single

commodity specialist — Kristin N. Sheaffer — and draws on submissions

from peer geological surveys, reported company data, trade statistics

compiled by Trade Data Monitor and reference series maintained by

industry associations.¹

The conservative bias has a specific purpose. The USGS data feeds into

US federal policy, including National Defense Stockpile procurement

planning and import-reliance calculations. A figure that overstates

global production can distort those calculations; a figure that

understates it is less problematic. Investors should therefore treat the

USGS number as a solid floor rather than a point estimate.

Metals Focus Methodology in Brief

Metals Focus is a subscription consultancy whose full Gold Focus report

is paid. Its open coverage — briefings published through the Silver

Institute and derivative reporting in specialised press — provides

visibility into its top-line figures and trend commentary.² The

methodology is built on primary channel checks: interviews with

producers, refineries, fabricators and bullion dealers, supplemented by

government statistics and trade-data reconciliation.

Because the consultancy serves clients who transact in physical gold and

need month-by-month granularity, its incentives favour a more inclusive

and operationally realistic count. That translates into a figure that

tends to sit above the USGS headline — but it also means Metals Focus'

number captures flows that have economic consequences (available supply,

refinery throughput, available inventory) even when they do not show up

cleanly in national statistics.

Why the Disagreement Matters

For investors, the gap between 3,300 and 3,672 tonnes has three

practical implications. First, any forecasting model that pegs future

supply to a base year inherits the assumption baked into that base. A

forecaster using the USGS starting point will produce a tighter supply

picture than one using Metals Focus, and the price implications differ

accordingly.

Second, country-level market share and political narratives depend on

the denominator. If Brazil produced 80 tonnes in 2025, it is 2.4 percent

of the USGS total but 2.2 percent of the Metals Focus total. Minor

arithmetic, but relevant when policy-makers cite rankings in negotiating

trade, investment or tariff terms.

Third, on the demand side, the 2025 record of 4,999.4 tonnes reported by

the World Gold Council needs to be reconciled against total supply,

which includes mine output plus recycling. The choice of mine-supply

estimate therefore affects how tight or loose the market looked through

the year — an input that matters to short-term price commentary.³

Out