The AI Copper Story Nobody Wants to Hear

By Siam Sukkhee Trading Co., Ltd — 2026-10-24 — Bing News — zinc price LME

I sat through enough commodity panels last week to know when a narrative is cracking. At Bloomberg's LME Week event in London on October 20, a room full of analysts from Macquarie, Mercuria and JPMorgan basically said the same thing: the AI copper story is mostly marketing.

Not that there's no copper demand from data centres. There is. But the sheer volume of capital flowing into AI infrastructure doesn't translate neatly into metal consumption.

Nicholas Snowdon, Mercuria's head of metals and mining research, described the AI capex cycle as "immense"—comparable to a US or China-scale investment programme. Yet metals demand had underperformed its historical relationship with industrial production and purchasing managers' indices. Something didn't add up. He put it plainly: "AI is not as metals-intensive as the headline capex numbers suggest." The money goes into warehouses and chips. Not copper piping.

Alice Fox at Macquarie pushed harder. She called the boom overhyped, full stop. Data centre investment, she said, was "really small" in the grand scheme of construction spending—just one of the few sectors still growing while the rest of the construction cycle stalled. But that's precisely the point. Growth at the margin gets all the attention. Weakness everywhere else gets forgotten.

Actually, let me correct that. The weakness isn't forgotten. It's actively ignored because it doesn't fit the bull case.

Gregory Shearer at JPMorgan took a different angle. He didn't expect data centres to be the game changer that sent copper into a million-tonne deficit. But he acknowledged something more subtle: "policies… are being made on a premise that whoever wins this race in AI controls the next 30 years geopolitically – and at the moment, copper is tied into that race." So you don't need to believe in the demand story to understand why prices are where they are.

That's the thing. US policy is the real driver here. Tariffs and stockpiling are propping up the price far more effectively than any fibre-optic-pulling-through-walls narrative could.

On the demand side, Snowdon said cyclical sectors—construction, housing, autos in China and Europe—are under pressure. Those have always been copper's bread and butter. Recovery probably needs interest rates to fall, which probably needs oil prices to come down first. Long chain of dominoes. That's real demand destruction, not a rounding error.

Snowdon called this "one of the greatest feats of policy support for a price in commodity market history." He wasn't celebrating.

What happens to prices if that tariff scenario plays out? They go up. Imports slow. The world outside the US becomes very tight. Demand gets choked off. Inflation pressure mounts. Everyone's happy except, you know, the people actually trying to use copper to build things.

The infrastructure story is real. Power grids, battery storage systems, renewable energy—that's all growing and copper-hungry. Shearer pointed out that these growth sectors have been insulating copper demand from the cyclical downturns elsewhere. But it's not the same as saying AI saved the metals complex. AI just happened to land when the structural bull case was already in motion.

When you strip away the headlines, what you're left with is this: policy risk is doing more work than demand forecasts ever could. The LME sees it. The traders see it. Everyone sees it. But the momentum machine needs a story that sounds less like "governments are artificially supporting prices through import restrictions" and more like "the future demands copper." One of those stories sells more conference passes.

So yes, there's copper demand from data centres and electrification. But no, it's not enough to explain where prices actually are right now. That's the uncomfortable bit. And it's the only bit that matters if you're trying to figure out what happens next.

Source: "A Bloomberg LME Week panel including Macquarie, JPMorgan and Mercuria analysts said US policy and constrained supply, not demand, drive copper." — Shanghai Metals Market

Tags: copper market Southeast Asia