Copper H2 2026: Supply Losses Are Real — the Question Is How Demand Responds

Key Assumptions: The Deficit Consensus Hardens

The remarkable shift in copper market forecasts over the past six months tells its own story. In October 2025, the International Copper Study Group (ICSG) projected a 209,000-tonne surplus for 2026. By April 2026, that surplus had been revised to roughly 96,000 tonnes. And by May, the ICSG flipped outright to a 150,000-tonne deficit — a swing of nearly 360,000 tonnes in a single forecast cycle. Financial institutions rarely converge on the same directional call with this degree of consistency. JPMorgan sees a deficit as large as 330,000 tonnes, driven by accelerating demand from AI data centers, grid electrification, and the energy transition. Only Goldman Sachs stands apart, forecasting a ~160,000-tonne surplus predicated on materially weaker Chinese demand.

The supply side of the ledger has deteriorated in ways that few predicted. Freeport-McMoRan delayed the full production restart at Grasberg in Indonesia — one of the world's largest copper deposits — from 2027 to 2028 following a September 2025 mudslide, then reduced its 2026 production guidance by approximately 35%. The Quebrada Blanca Phase 2 ramp in Chile has absorbed billions in capital without achieving nameplate throughput. Treatment and refining charges (TC/RCs) remain at or near record lows, reflecting an extraordinary squeeze in concentrate supply — smelters are competing ferociously for every available tonne of feed. These are not cyclical signals. They are structural.

On the demand side, the contours are more contested. Global refined copper use grew at a subdued pace through early 2026, with the ICSG revising its 2026 growth forecast down to 1.6% from an earlier 2.1%. China, which consumes more than half the world's copper, saw demand contract -8% in Q4 2025 — a sharp deceleration that continues to reverberate through market psychology. But looking forward, the demand picture fractures along sector lines. Traditional construction and manufacturing remain soft, while AI/data-center infrastructure, grid modernization, and EV production are absorbing copper at an accelerating rate.

Trigger Variables: The Five Levers That Will Decide H2

Five variables will determine which price path copper follows through the second half of 2026. They form the decision framework for the scenarios below.

1. Chinese Demand Recovery

The single biggest variable. China's property sector remains depressed, and the Q4 2025 demand drop of -8% was the steepest quarterly contraction in years. However, Chinese refined copper imports reached 452,000 tonnes in April 2026, suggesting that restocking and infrastructure spending are providing a floor. A stabilization — let alone a recovery — in Chinese copper demand would profoundly tighten the global balance.

2. Section 232 Tariff Decision

The United States is expected to decide by late June 2026 whether to impose Section 232 tariffs on copper metal imports. Since the probe was ordered in February 2025, Comex copper stocks have surged 550% as traders front-run the decision. As recently as May 22, Trafigura began withdrawing large volumes of copper from LME warehouses in New Orleans, effectively locking in pre-tariff pricing. The ruling — whatever its form — will ripple through global pricing and arbitrage flows for the rest of the year.

3. Mine Supply Disruptions

Beyond Grasberg, the pipeline of new copper supply is dangerously thin. No major greenfield projects are scheduled to reach first production before 2029–2030. Existing operations in Chile and Peru — the world's two largest producing countries — face declining ore grades, water constraints, and political headwinds. Wood Mackenzie projects that global copper demand will rise ~24% by 2035, yet the mine-development pipeline can barely sustain current output, let alone absorb that growth.

4. AI and Data-Center Demand Acceleration

JPMorgan estimates that AI-related data centers could account for roughly 475,000 tonnes of copper demand in 2026, an annual increase of about 110,000 tonnes. This is not theoretical — hyperscalers are placing multi-year copper procurement contracts, and the electrical infrastructure inside a single large data center can consume thousands of tonnes of copper cable and busbar. As the AI buildout intensifies, this demand stream becomes increasingly price-inelastic.

5. Energy Transition Infrastructure

Grid modernization, renewable-energy deployment, and EV manufacturing together form a structural demand super-cycle. Copper intensity per EV is roughly 3–4x that of an internal-combustion vehicle, and every gigawatt of offshore wind requires approximately 8,000 tonnes of copper. These are long-duration demand drivers that do not respond sharply to near-term price signals — they are policy-backed and capital-committed.

Three Scenarios for H2 2026

Base Case: Orderly Deficit, Gradual Price Re‑rating ($12,000–14,000/mt)

Probability: ~50%. In this scenario, Chinese demand stabilizes at current levels — no recovery, but no further deterioration. The ICSG deficit of 150,000 tonnes proves broadly accurate, with supply losses concentrated in the first half of the year. Section 232 tariffs are imposed in a moderate form, creating a persistent but manageable premium in US markets. AI/data-center and energy-transition demand continues its scheduled ramp. LME copper drifts higher through Q3, breaching $12,000 by August, and trades between $12,000–14,000 in Q4. Refined copper imports into the US surge as buyers lock in supply ahead of tariff implementation. This is the path of least resistance: not a price explosion, but a steady grind higher as the market slowly prices in structural scarcity.

Bull Case: Supply‑Shock Pricing ($14,000–16,000/mt)

Probability: ~25%. A single catalyst — or, more likely, a cascade of triggers — pushes prices toward the all-time highs set in May 2024 and beyond. A further major supply disruption at a large mine complex (Chuquicamata, Escondida, or Grasberg) removes 50,000–100,000 tonnes from the annual balance. Chinese demand posts a modest recovery of +2–3% in H2, reversing the Q4 2025 contraction. Section 232 tariffs come in at the aggressive end of expectations, physically diverting metal away from China and into the US and creating a bifurcated global market. TC/RCs, already at historic lows, go negative in selected term contracts — signaling that smelters are paying miners to take concentrate. LME copper crosses $14,000 by September and tests $15,000–16,000 by year-end. Buyers who did not lock in term contracts face spot premiums of $500–800/mt above LME. This scenario mirrors the 2024 rhodium-style asymmetry: in a deficit market, prices do not just rise — they gap.

Bear Case: Demand Destruction ($10,000–12,000/mt)

Probability: ~25%. Chinese demand remains weak through H2 2026, with Q3 showing another quarter of contraction. The property-sector downturn deepens, and industrial production fails to rebound. Goldman Sachs' surplus forecast of ~160,000 tonnes proves more accurate than the deficit consensus. Section 232 tariffs are either delayed or imposed in a weak form that fails to disrupt trade flows. AI/data-center capital expenditure — while still growing — decelerates from exponential to linear as hyperscalers face their own return-on-investment questions. LME copper drifts back toward $10,000, with Q4 averaging in the $10,000–12,000 range. In this world, the structural bull case is not invalidated — merely deferred. The deficit arrives in 2027 instead of 2026. Buyers who need copper through 2027–2028 should use any pullback below $10,500 to layer in forward coverage.

Decision Matrix

Scenario	Prob.	Price Range	Key Signal to Watch
Base Case	50%	$12,000–14,000	Chinese imports stabilize; TC/RCs bottom
Bull Case	25%	$14,000–16,000	Major mine outage; China returns to growth
Bear Case	25%	$10,000–12,000	China demand contracts further; tariffs delayed

The base case — orderly deficit, gradual price appreciation — carries the highest probability. But the asymmetry of outcomes bears emphasis: the bull case implies a +40–57% move from current levels, while the bear case implies only a 0–+18% range. The distribution is right-skewed. Copper's path of least resistance is higher, unless Chinese demand actively deteriorates.