Countries that buy antimony often record far more of it than their suppliers register as sent. Fiftyone traced the gaps across 13 countries. Most have ordinary explanations, a few do not, and one undermines the most widely cited figure for US demand.
Italian customs recorded 930 tonnes of antimony oxide of Chinese origin arriving in 2025. China's customs data, as compiled by the International Trade Centre, records no antimony oxide exported to Italy that year at all.
Italy is not an isolated case. Fiftyone compared what exporters list as shipped with what importers log as received, for antimony oxide, metal and ore, across 13 countries in Asia, Europe and the Americas. Where trade is direct, with no third country involved, the figures Fiftyone checked were broadly in line. Where it is not, the two sides can be far apart, with one sometimes recording nothing at all, and the gaps follow a small number of patterns.
The US figures
By weight
Trade, year | Exporter's record | |
China to the US, oxide, 2025 | 2,115 t | |
Belgium and France to the US, oxide, 2025 | 19,567 t | |
South Korea to the US, oxide, 2025 | 1,768 t | |
Mexico to the US, oxide, 2025 | 1,061 t | |
Bolivia to the US, oxide, 2025 | 2,716 t | |
India to the US, oxide, 2025 | 1,828 t | |
Thailand to the US, oxide, 2025 | 4,221 t |
Measured in tonnes, the US records far more oxide than its suppliers say they shipped, roughly two to three times as much for most of them. In dollars, the picture changes. For most suppliers the US value and the exporter's value broadly agree, which points to the same shipments recorded at the wrong weight rather than extra oxide arriving unrecorded. Thailand runs the other way: both sides record roughly the same tonnage, but the US records only a fraction of Thailand's value.
By value
Trade, year | Exporter's record | |
China to the US, oxide, 2025 | $11.5m | |
Belgium and France to the US, oxide, 2025 | $301.3m | |
South Korea to the US, oxide, 2025 | $44.7m | |
Mexico to the US, oxide, 2025 | $6.1m | |
Bolivia to the US, oxide, 2025 | $62.2m | |
India to the US, oxide, 2025 | $33.5m | |
Thailand to the US, oxide, 2025 | $14.3m |
The likeliest cause is the way the US Census Bureau processes import records. Before publishing, its system tests each shipment's price per tonne against a range set from up to five years of past data, and automatically corrects records that fall outside it. When antimony prices tripled after China's controls, genuine shipments priced at $40,000 a tonne or more would have looked like errors against ranges built on far lower prices. If the system kept each declared value and recalculated the weight, the result would be what the tables show: values that match the exporters' and tonnages that do not. The recorded prices point the same way. Between August 2024 and April 2025, oxide from China, Belgium, France, Mexico, South Korea, India and Bolivia was each recorded at almost exactly $4,500 a tonne in at least one month, far below the market and identical across unrelated suppliers. By mid-2025, as the ranges caught up with the new prices, recorded prices returned to market levels. It remains an inference from the data.
That matters beyond trade statistics. The US Geological Survey puts US apparent consumption of antimony at 45,000 tonnes in 2025, up from 28,600 tonnes in 2024. That figure includes 39,000 tonnes of antimony in oxide imports. Antimony trioxide, the main oxide traded, is about 83.5 per cent antimony by weight, so that implies about 47,000 tonnes of oxide. The 47,000 tonnes closely matches US customs records for January to August 2025, including the inflated months, scaled up to a full year. Priced at market rates, the $524m of oxide the US imported across 2025 suggests 12,000 to 15,000 tonnes, against 35,945 recorded. The most cited figure for US antimony demand is most likely too high.
Where they agree
The data is not simply broken. Germany recorded 20 tonnes of Chinese-origin oxide in 2025, and China recorded 20 tonnes sent to Germany. Japan recorded 103 tonnes of Chinese antimony metal, and China recorded 100. Before export controls, the US and Chinese records of oxide trade between them matched within 4 per cent, in both 2022 and 2023. Where antimony travels in one step from seller to buyer, both sides count it the same way.
The gaps appear when it does not. Five reasons account for most of them.
Where the records disagree
Where exporters record less
Trade, year | Importer's record | Exporter's record |
Oxide | ||
China to the EU, oxide, 2025 | ||
China to South Korea, oxide, 2025 | ||
Metal | ||
China to South Korea, metal, 2025 | ||
Vietnam to its main buyers, metal, 2025 |
One: origin is not the same as route
Exporters record where goods are sent. Importers outside the EU, and the EU itself for goods arriving from outside the bloc, record where goods were made. A consignment of Chinese oxide shipped to a trader in a third country and sold on to Italy appears in China's records under the third country and in Italy's records as Chinese.
Italy also acts as an entry point for the rest of Europe. It took in 3,118 tonnes of antimony oxide in 2025 and sent out 952 tonnes. About 70 per cent went to other EU states, led by Germany, Belgium and Poland, and Tunisia was the largest buyer outside the bloc. Once goods are inside the EU, movements between member states carry no origin, so Chinese oxide that enters through Italy and is used in Germany is counted as Chinese only once, in Italy. Italy accounted for about four-fifths of the EU's 1,184 tonnes of Chinese-origin oxide imports in 2025.
That explains why Italy dominates the EU figures. It does not explain why China's records show nothing going there, and the mismatch is not new: in 2023, before any controls, Italy recorded 184 tonnes of Chinese oxide against 43 tonnes in China's records. Chinese oxide also arrived in Italy at €21,100 a tonne in 2025, against €35,300 for South Korean oxide, making it Italy's cheapest large source of supply.
Two: hubs
Thailand handled almost no antimony ore until 2024. Its own customs records show ore imports rising from 76 tonnes in 2023 to about 69,000 tonnes in 2025, more than nine-tenths of it from Myanmar. Its metal imports rose from $33m to $460m, mostly from Myanmar and Vietnam. In 2025 Thailand exported $270m of metal, including $97m to Belgium and $62m to Japan, and about 4,500 tonnes of oxide, 3,800 of it to the United States.
Importers disagree about where that material comes from. The EU records metal from Thailand as Vietnamese or Myanmar origin. Thailand records 3,600 tonnes of metal shipped to Japan, while Japan records 427 tonnes of Thai-origin metal and 1,700 tonnes of Myanmar-origin. South Korea's record matches Thailand's by weight, 845 tonnes against 867, but Korea values it at $46m and Thailand at $12m. The US records fewer tonnes than Thailand, 1,589 against 2,400, but a higher value, $73.8m against $44.5m. The same metal is counted under three different origins, and at very different prices, depending on the importer.
Thailand's own metal figures stopped adding up during the boom. In 2023 its metal trade was priced at the market, about $11,500 a tonne. By 2025 its main flows were valued at $13,000 to $21,000 a tonne, well under half the market price, so its tonnages and values for that year cannot both be right.
Three: classification
Vietnam's statistics show only 614 tonnes of unwrought antimony exported in 2025, under code 8110.10. Vietnam's buyers recorded about 3,860 tonnes: Japan 847, the EU 2,626, South Korea 126 and the US 265. Vietnamese shipment records explain the gap. Its smelters declare ingot under 8110.90, the catch-all heading for other antimony articles, where Vietnam reports a further 4,300 tonnes of exports, much of it to Thailand, while importers record the same ingot as unwrought metal. Counting both codes, Vietnam's exports to Japan come to 813 tonnes against Japan's 847.
Classification gaps of this size are not unique to antimony. Italy recorded 1,371 tonnes of titanium oxide from China in 2025. China recorded 2 tonnes sent to Italy under that code, a gap consistent with China booking most titanium dioxide under a different heading, such as pigments.
Classification can also be deliberate. In December 2025 a court in Shenzhen jailed 27 people for smuggling more than 166 tonnes of antimony ingots out of China in February and March that year through concealment, disguise and false declaration. Customs seized about 96 tonnes. A trader quoted by Reuters described shipments relabelled as iron, zinc or art supplies. Antimony that moves this way appears in no country's antimony statistics.
Four: conversion
When antimony changes form, its country of origin can change with it. Under most countries' customs rules, turning metal into oxide creates a new product, which takes the origin of the country where the conversion happened. South Korea imported about 1,500 tonnes of antimony metal in 2025, mainly from Thailand and China, and exported about 2,700 tonnes of oxide, roughly 1,000 tonnes of it to the United States. Korea Zinc recovers antimony as a by-product at its Onsan smelter, a record 971 tonnes in the first quarter of 2025 and about 10 tonnes a day by March 2026, which closes the Korean balance.
Chinese metal turned into oxide in Korea leaves as Korean origin. That is lawful, and it means this antimony can reach the United States as Korean oxide without appearing as Chinese in anyone's statistics. Oxide made in China is different: it keeps its Chinese origin wherever it travels, which is why it still shows up in US records.
Five: bad data
Sometimes the figures are simply wrong. The US oxide records set out above are the clearest case: Belgium and France's export figures imply €44,000 to €49,000 a tonne (about $50,000 to $55,000), in line with the market, while the US records of the same trade imply $14,000 to $19,000.
What the timing shows
The US recorded $11.5m of Chinese-origin oxide in 2025, though China records nothing sent to the US after September 2024. Of that, $9.5m arrived between January and April 2025. US general import figures, which count goods as they enter the country, match imports for consumption exactly, so this was not stock released from US bonded warehouses. It entered the US in 2025, most likely material bought or held outside China before the controls.
On 9 May 2025, China's export-control coordination office announced a campaign against “false declarations, concealed smuggling, and ‘third-country’ re-exports” of strategic minerals including antimony, accusing overseas entities of colluding with people inside China to evade the controls. Within weeks, Chinese-origin oxide arriving in the US fell from $3.0m in April to $0.8m in May and $0.1m in June. Mexican-origin oxide followed the same arc. The US recorded almost none before November 2024, the month before China banned exports to the US. From then it arrived every month, worth between $0.8m and $1.9m, until a last $0.8m in May 2025. After that, none. The sequence fits Reuters' reporting that Chinese antimony reached the US through Thailand and Mexico. It does not prove it.
Stockpiles
Trade figures only count antimony that moves. Some of it is sitting still, in company and state stockpiles. When the Fanya Metal Exchange collapsed, China Minmetals bought its entire 18,661 tonnes of antimony in 2019, about a fifth of China's annual output at the time. What has happened to that stock since is not on the public record. Hunan Gold ended 2024 with 3,742 tonnes of antimony products in stock, up 215 per cent. Huayu Mining's Tajik joint venture, Talco Gold, ended 2025 holding 2,995 tonnes of antimony in concentrate, having sold 122 tonnes of the 2,338 it produced.
China's own balance points the same way. It imported 59,348 tonnes of ore and concentrate in the first half of 2026, more than in the whole of 2025. Even if that ore averaged only 10 per cent antimony, it would hold about three times the antimony in the 2,350 tonnes of oxide China exported over the same six months, and China exported no metal at all. Demand at home was weakening. Chinese financial media have reported a 25,000-tonne state reserve, but neither Huayu Mining nor Hunan Gold, the two companies named as its first storage units, mentions one in its 2025 annual report, and Fiftyone found no official document confirming it.
Outside China, new smelters in Cambodia and Vietnam expanded during the 2025 price spike. Jiu Sheng Non-ferrous Metals, on the Thai-Cambodian border, advertises capacity of 600 tonnes of ingot a month. Market commentary describes smelters and traders sitting on unsold metal.
What the figures leave open
Three gaps remain unexplained. Only Italian customs, which holds the country of consignment for each declaration, can show how Chinese-origin oxide reached Italy. Thailand reports $106m of oxide exports to the US in 2025, against $14.3m of Thai-origin oxide in US records, though both sides record roughly 4,000 tonnes. US Customs and Border Protection alone can say why the declared value is a seventh of Thailand's. And a small, steady flow of low-priced Chinese-origin oxide was still entering the US in 2026.
For buyers and for defence planners, the lesson is that trade statistics record where antimony was last declared, not where it was mined. The official figures understate how much Chinese material still reaches the market and how far it relies on Southeast Asian intermediaries.
In the United States, they also overstate how much antimony the country uses. Strip out the inflated oxide records and the US Geological Survey's 45,000 tonnes for 2025 falls to something closer to the 20,000 to 28,000 tonnes it recorded before China's controls, if not lower. For anyone sizing the American market, or the shortfall that new mines and refineries are being funded to close, the lower figure may be the more realistic starting point.
