Copper at $14,000: the AI trade is rewriting the metals chart — and the Global South wants a seat at the table

Copper closed above $14,000 a metric ton on the London Metal Exchange on 29 June 2026, according to a Reuters wire circulated the same afternoon, the latest leg of a rally that has carried the red metal from a five-year average closer to $9,000 to record territory in under eighteen months. The trigger, on this occasion, is the same one traders keep reaching for: artificial intelligence. Reuters reports that demand from AI data centres, plus renewed interest in copper as an inflation hedge, has dragged a plumbing-and-wires industrial commodity into the same portfolio conversation as Nvidia and the hyperscalers.

That framing is half right and half dangerous. Copper is not a tech stock. It is dug out of the ground in Chile, Peru, the Democratic Republic of the Congo and, increasingly, recycled in China; it is smelted in places that have spent two decades coping with the boom-and-bust cycle of a metal the West treats as decorative until it suddenly isn't. Treating copper as a hedge-and-AI play is fine for a Bloomberg terminal. It is also the setup for a much bigger argument about who captures the value of the energy transition when the metal underpinning it trades at all-time highs.

The chart tells one story; the mine tells another

The Reuters dispatch is precise about the demand side. AI data centres are power-hungry; power transmission is copper-intensive; grid build-outs, particularly in the United States and the European Union, are pulling tonnage the market did not budget for. On top of that, fund flows have treated copper as a macro hedge against the inflation that tariffs, fiscal expansion and energy-price volatility keep threatening. The combination has pushed inventories on LME warehouses down and the price up.

The supply side is messier. The largest copper mines on earth sit in jurisdictions that the same Western investors now buying copper-ETF exposure do not especially want to underwrite politically. Chile's Codelco and BHP's Escondida carry decades of royalty disputes and water-rights litigation. Peru's Las Bambas has been a stop-start operation owing to community blockades. The DRC's copperbelt, the fastest-growing source of new supply, ships concentrate that is overwhelmingly routed through Chinese smelters — a structural fact that has very little to do with the AI narrative and everything to do with who built the processing capacity the West did not.

The China file the wire won't write

Here is the part that does not make the front page of the financial press. China is already the world's largest copper consumer and the largest refiner of concentrates, including those pulled out of African and South American mines that Western capital helped finance. Reuters notes the AI demand story; Reuters is less interested in the fact that the refining capacity required to convert Chilean and Congolese concentrate into the wire bar, rod and foil the data-centre builders want is overwhelmingly Chinese. CATL's copper-foil subsidiary, listed separately on the Shenzhen exchange, has expanded capacity through 2025 at a pace no Western fabricator has matched.

The structural argument is plain. Industrial policy coherence matters. For two decades, Chinese state and provincial credit has underwritten mid-stream processing capacity — smelters, refiners, foil plants — on the assumption that copper demand would rise. That bet is paying off at $14,000. The Western equivalent, post-IRA and EU Critical Raw Materials Act, is only just starting to be financed; the first new Western copper smelter in a generation is years from first pour. The result is that the price discovery may happen in London, but the value-add chain runs through Tianjin, Ningde and Huizhou.

The Global South's leverage — and its limits

This is not a story the major miners want to tell, but it is one that the producing countries are increasingly willing to. Chile's state-owned Codelco has spent the last two years renegotiating royalty frameworks and pushing harder on lithium alongside copper. Peru's government has used community-relations crises at MMG's Las Bambas as leverage on tax terms. The DRC, which produces more than 70% of the world's cobalt and a growing share of copper, has been tightening conditions on artisanal supply chains and on the Chinese offtake arrangements that determine where the concentrate ends up.

The leverage is real, but it is bounded. Copper is a globally priced commodity; no single producer can set the LME benchmark. The producing countries can capture more rent through royalties and processing requirements — Zambia's push to mandate in-country cathode production is the model — but they cannot manufacture demand that doesn't exist. And here the AI narrative cuts both ways: if hyperscaler capex slows, the marginal buyer disappears, and the boom looks, in retrospect, like the dotcom-era fibre build-out all over again.

Stakes, in plain prose

If the AI-driven demand story holds, copper has further to run, the data-centre build-out accelerates the West's reliance on imported metal and imported refining capacity, and the producing countries — Chile, Peru, the DRC, Zambia, Indonesia, Mongolia — sit on a more valuable asset than they did a year ago. The political consequence is that resource nationalism becomes harder for Western capitals to dismiss as extractive populism; it starts to look like ordinary bargaining over rents the market has handed the producers.

If the AI demand narrative breaks — if model-efficiency gains collapse the data-centre capex curve the way fibre overcapacity hit the telecoms build — the chart snaps back, the inventories refill, and the producers are left holding concessions sold at peak prices. That is the historical pattern. It is the part of the story the inflation-hedge framing is designed to obscure.

The honest read is that $14,000 copper is a market price plus a belief. The belief is that AI will keep building, the grid will keep expanding, and the metal will keep being in the wrong place at the right time. Whether that belief deserves the multiple it is currently being given is a question for the people writing the cheques, not the people selling the ore.

Monexus framed this as an industrial-policy story with a commodities wrapper, not the other way round. The Reuters wire on 29 June 2026 supplied the price; the structural reading on refining capacity, producer leverage and the China-processing chain is this publication's own analysis and should be read as such.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://x.com/reuters/status/2071632474120286208
• https://x.com/sknerus_/status/2071376978435883008