Here’s the thing nobody wants to admit: the copper market is entering 2026 with a supply-demand mismatch that can’t be solved by throwing money at it.
You can’t disrupt geology. You can’t software-engineer a new mine into production. And you definitely can’t build out AI infrastructure, renewable capacity, and grid modernization simultaneously without hitting a wall.
That wall is copper. And we’re about to hit it hard.
The Deficit Nobody Agrees On
Let’s start with the numbers, because even the experts can’t agree on how bad this gets.
J.P. Morgan Global Research forecasts a deficit of roughly 330,000 metric tons in 2026. The International Copper Study Group (ICSG) is more conservative at 150,000 metric tons. Goldman Sachs: in a move that’s either contrarian brilliance or willful optimism: is calling for a 300,000 metric ton surplus.
That’s not a rounding error. That’s three different narratives about the same market.
The disagreement comes down to demand elasticity. Goldman believes higher prices will throttle consumption and boost recycling fast enough to clear the market. JPMorgan and ICSG assume demand stays sticky: that buyers will pay up rather than walk away.
Here’s where it gets interesting: both camps are making bets on human behavior, not just supply curves. And in a market driven by policy mandates (electrification targets, data center build-outs, defense spending), demand doesn’t respond to price signals the way it used to.
The Demand Side: Three Engines Firing at Once
Copper demand in 2026 isn’t coming from one source. It’s coming from three simultaneous build-outs that don’t defer easily.
Energy transition infrastructure is the biggest driver. Electric vehicles, battery storage, renewable capacity, and transmission grids are expected to add more than 7 million metric tons of annual demand by 2040, according to S&P Global. That’s not aspirational: it’s already baked into government policy and corporate capex plans.
AI and data centers represent the new wildcard. Hyperscalers are hammering out deals for nuclear power and building facilities that consume copper at rates the grid wasn’t designed to handle. Each large-scale AI data center can require 1,500 to 2,000 metric tons of copper. Per facility.
That’s not a typo.
Traditional demand: construction, power generation, industrial equipment: still accounts for 53% of total consumption and will hit 23 million metric tons by 2040. This is the base load. It doesn’t go away just because Tesla and Microsoft need more wire.
They’re all competing for the same metal. And they’re all pulling from a supply base that’s struggling to keep up.
The Supply Side Reality Check
Here’s where the numbers get grim.
Global copper production is projected to peak in 2030 at 33 million metric tons, then decline. Meanwhile, demand is expected to reach 42 million metric tons by 2040. That’s a 10 million metric ton structural deficit: roughly equivalent to wiping out Chile’s entire annual output.
Even if recycled copper scrap more than doubles from 4 million to 10 million metric tons by 2040, it doesn’t close the gap.
The brutal reality: overcoming this deficit requires an additional 10 million metric tons of primary supply capacity by 2040. Absent significant new investment, global primary supply could actually produce 1 million metric tons less than it does today.
And 2026 isn’t immune to this squeeze. Freeport-McMoRan’s Indonesian mine accident will result in approximately 500,000 tons of lost copper over 12 to 15 months, with phased restart targeting only 85% of normal capacity by mid-2026. The Kamoa-Kakula complex in the Democratic Republic of Congo delayed its 500,000-ton annual target from 2026 to 2027.
Those aren’t forecast risks. They’re already happening.
What This Does to Price
The market is pricing in the deficit. Spot copper hit an all-time high of $13,300 per metric ton on January 6, 2026. That’s not speculative froth: it’s supply and demand doing what they do.
J.P. Morgan expects average prices around $12,075 per metric ton in 2026, with a Q2 peak near $12,500. Citigroup is even more aggressive, suggesting $13,000 to $15,000 if shortages persist.
But here’s the kicker: Goldman Sachs expects prices to decline later in 2026 as high prices dampen demand and lift scrap supply, particularly given weakened Chinese copper consumption.
So which is it?
The truth is probably somewhere in the middle. Prices will spike when physical tightness hits specific markets (North America, Europe), then soften when buyers defer non-critical purchases. But the structural deficit doesn’t go away. It just gets deferred, restocked, and repriced.
The Investor Playbook: What Actually Matters
If you’re allocating capital in this market, here’s what separates signal from noise.
Primary producers with low costs and reserve life win. Companies that can deliver copper at cash costs below $2.50 per pound have pricing power. Names with 20+ years of reserve life aren’t just selling metal: they’re selling optionality in a structurally tight market.
Brownfield expansions beat greenfield development. Permitting timelines and capital intensity favor projects that can add capacity to existing operations. Greenfield mines face 7 to 15 year development cycles. The market doesn’t have that kind of time.
Recycling and secondary supply players get a bid. Scrap processors, urban mining initiatives, and companies with closed-loop supply chains will see margin expansion as primary-secondary spreads widen.
Exposure to policy-driven demand is sticky. Copper consumed in grid infrastructure, EV charging networks, and renewable capacity isn’t discretionary. It’s mandated. That demand doesn’t evaporate when prices spike: it just gets financed differently.
Geopolitical risk is now a valuation input. Resource nationalism, export controls, and supply chain concentration (particularly around Chinese smelting and refining) create risk premiums that didn’t exist five years ago. Jurisdiction matters more than grade in some cases.
The Uncomfortable Truth
Here’s what makes this particularly nasty for investors: the copper deficit isn’t a cyclical problem that resolves with a recession or demand destruction. It’s structural, policy-driven, and locked in by commitments that governments and corporations can’t easily walk back.
The grid needs to be rebuilt. AI infrastructure is getting built regardless of copper prices. Defense spending: a quiet but meaningful demand driver: isn’t slowing down.
And the mining industry? It’s dealing with grade decline, permitting delays, capital discipline, and ESG pressures that make rapid supply responses almost impossible.
The clock is already ticking. New mining capacity takes a decade to permit and build. The demand is hitting now.
That’s the structural mismatch investors need to price. Not “if” there’s a deficit, but “how long” it lasts and “who” captures the margin when supply finally catches up: if it ever does.
Welcome to the new reality. There’s not enough to go around.
