THE PHYSICAL FLOW LEDGER
Six official series, one scale, no manufactured balance
Compare upstream production, refinery output, border flows and reported use without counting the same metal twice or treating a survey measure as an accounting identity.
Accepted public-domain release
USGS MCS 2026 v1.3 · 2021–2025 · 2025 estimated · normalized history served locally; the official source is used only for offline update checks.
All six lines use the same physical scale. Move the pointer, tap or drag, or focus the chart and use Left, Right, Home and End. Exact values remain below the chart.
EXACT-YEAR EVIDENCE
2025 estimate: do not add these rows into one total
Primary refinery output can process already-counted domestic or imported feed. Reported consumption is a different survey measure.
Arithmetic trade gap: −60 tExports minus imports for consumption. This sign does not identify where exported metal was mined, refined, previously held or ultimately used.
Open the complete five-year source table
| Year | Mine | Primary refinery | Secondary refinery | Imports for consumption | Exports | Reported consumption | Exports − imports | Status |
|---|---|---|---|---|---|---|---|---|
| 187 | 181 | 92 | 192 | 386 | 265 | +194 | Published | |
| 173 | 181 | 93 | 138 | 420 | 252 | +282 | Published | |
| 170 | 177 | 96 | 215 | 252 | 253 | +37 | Published | |
| 163 | 180 | 89 | 190 | 289 | 210 | +99 | Published | |
| 160 | 170 | 90 | 320 | 260 | 150 | -60 | Estimated |
Interpretation boundary: these series do not measure total U.S. gold demand, beneficial ownership, retail bullion demand, ETF flows or Treasury transactions. A flow difference is not evidence of scarcity, fair value, future price direction or a trading signal; this chart is not a price forecast.
THE DIRECT ANSWER
U.S. gold trade changed direction inside this five-year release
Exports exceeded imports for consumption from 2021 through 2024, with the largest arithmetic gap at 282 metric tons in 2022. In the 2025 estimate, imports rose to 320 tonnes while exports were 260 tonnes, reversing the sign to -60 tonnes.
That reversal is a border-flow observation, not proof of a shortage or a change in U.S. beneficial ownership. The trade rows span ores and concentrates, dore and bullion, and exported material can have a different origin or prior processing path.
WHY THE LINES DO NOT ADD UP
Refinery output is transformation; reported consumption is a separate measure
Mine production
Domestic gold content recovered from mines. It excludes imported feed and secondary scrap.
Refinery production
Primary and secondary refinery output describe processing channels. Adding primary refinery output to mine output can double count metal.
Reported consumption
A separate source series, not a residual created by production plus imports minus exports.
For this reason, GoldObserve shows aligned lines and an exact-year ledger rather than a stacked total or a bridge forced to zero.
SCOPE BEFORE STORY
What the five-year ledger can and cannot answer
It can compare the scale and direction of six published annual physical-flow measures inside one MCS edition. It cannot reconstruct monthly customs flows, partner countries, beneficial owners, inventories outside the Treasury line, ETF activity, retail bullion purchases or the final use of traded metal.
The separate Census HS 7108 candidate remains release-blocked after its official bulk endpoint returned HTTP 403 in the 2026-08-15 prototype and normalized CSV rights remained unresolved. This page does not silently substitute for that dataset.
SOURCE IDENTITY
One public-domain edition, retained with file hashes
The source is the USGS MCS 2026 Gold chapter, version 1.3. The retained chapter identity is 138,864 bytes with SHA-256 c2fef62f665d3334302b8b9bd32e2da6f40b3f8136d1ae00ffa102e196943627.
The companion MCS2026_Commodities_Data.csv is 3,189,010 bytes, MD5 36185ff3742087e1dd90c52fe634fe12 and SHA-256 582a0aa231aea53d8a97dc8d1cd3dfa5f885cf3760353e3d029d7f0ae4fbaaf5. The 2025 row remains visibly estimated.
CONTINUE THE SUPPLY RESEARCH
Use the page that matches the evidence question
QUESTIONS
U.S. gold supply-flow FAQ
Is this a complete U.S. gold supply balance?
No. The USGS Gold chapter reports distinct production, refinery, trade and use measures. Primary refinery output can process domestic or imported feed, and reported consumption is not an apparent-consumption identity. GoldObserve does not force the rows to reconcile.
What do imports and exports cover?
The MCS Gold chapter's trade note covers gold content in ores and concentrates, dore and bullion. The rows do not identify beneficial ownership, final use, investment demand or every separately classified gold product.
Why is primary refinery output separate from mine production?
Refining transforms feed into commercial-grade metal. The feed can include domestic mine material or imported material, so adding refinery output to mine output would risk counting the same metal twice.
Does reported consumption equal total U.S. gold demand?
No. It is the source's reported-consumption series and does not include every investment or ownership flow. It should not be relabelled total demand.
Do these flows predict the gold price?
No. Annual physical-flow observations do not by themselves determine scarcity, fair value or future price direction.