Goldman vs. J.P. Morgan: Copper’s 2026 Battle—Surplus or AI-Driven Structural Deficit?

The outlook for copper in 2026 is defined by a stark clash between near-term forecasts and emerging structural realities. At the center of the debate is Goldman Sachs, which sees a clear path for prices to fall. The bank's base case is for a 300,000-ton surplus in the global market this year, driving the London Metals Exchange (LME) copper price to decline to $11,000 per tonne by the end of 2026. A key near-term driver for this forecast is a significant pullback in demand from China, where refined copper demand fell to -8% year-on-year in the fourth quarter of 2025. This slowdown, following the waning of stimulus and tariff-related front-loading, is creating the surplus that Goldman anticipates will pressure prices later in the year.

Yet this near-term surplus narrative is rapidly being challenged by a more bearish structural outlook. The International Copper Study Group (ICSG) has reversed course, now officially forecasting a 150,000-metric-ton deficit for 2026. Even steeper is J.P. Morgan's projection, which calls for a 330,000-metric-ton shortfall. The divergence hinges on the pace of new demand versus supply. While Goldman's model assumes a manageable surplus, the other analysts point to a powerful new source of demand: artificial intelligence. J.P. Morgan's deep deficit is driven heavily by the massive material demands of new hyperscale data centers, with estimates suggesting these facilities alone could consume hundreds of thousands of tons.

This sets up a tense year for copper. The dominant near-term pressure will be the expected surplus, which could keep prices capped in the $10,000-$11,000 range. But the market's path is unlikely to be smooth. The sharp correction from record highs earlier this year shows how quickly sentiment can shift. If structural supply constraints-like prolonged mine closures and downgraded output guidance-materialize faster than anticipated, or if AI demand accelerates, the deficit forecasts could quickly become reality. The year's price action will likely face downward pressure, but it may be interrupted by a sharp correction if the market's structural deficit narrative proves more powerful than the surplus forecast.

Inventory Dynamics: The US Stockpiling Hangover

The recent rally that pushed copper to record highs was powered by a potent mix of supply shocks and speculative front-running. The initial surge was fueled by the 22% jump from under $11,000 to a record $13,387 in early January, driven by concerns over the Grasberg mine and geopolitical tensions. But the rally's most significant amplifier was a wave of panic-buying ahead of anticipated US trade restrictions. This created a temporary scarcity premium outside the United States, as buyers rushed to secure metal before potential tariffs hit.

That dynamic is now breaking down. The primary catalyst for the initial rally-a 15% US tariff on refined copper-remains uncertain, with Goldman's base case pointing to a mid-2026 announcement. Yet the market's reaction to this uncertainty has already played out. The speculative front-running has led to a massive build in US inventories, with US COMEX inventories reaching a 40-year high of 590,000 tonnes. This "de facto stockpiling" has saturated the domestic market, removing the key near-term price support that was created by the fear of a refined copper shortage.

The hangover is clear. As US stocks continue to rise, the artificial scarcity premium evaporates. Goldman Sachs notes that with the "stockpiling fever" of late 2025 has broken, the market is entering a normalization phase. The immediate pressure on prices is shifting from a scarcity narrative to a surplus one, as the excess inventory built during the tariff anticipation period now sits as a tangible overhang. This inventory build is a critical factor that Goldman's bearish forecast must account for, as it directly contradicts the tightness that fueled the January peak.

Structural Demand vs. Cyclical Supply: The Long-Term Bet

The long-term bet for copper hinges on a fundamental imbalance between two powerful forces. On one side is a surge of structural demand, driven by the electrification of everything and a new, concentrated source of need: artificial intelligence. On the other is a supply chain that continues to stumble, with global production hampered by operational failures and downgraded guidance. The market's path in 2026 will be the battleground where these forces collide.

The demand side is now undeniable. J.P. Morgan's projection of a 330,000-metric-ton shortfall is the most extreme, and it is heavily weighted by AI. The bank estimates data centers will consume roughly 475,000 tons of copper this year. Each new hyperscale facility, housing power-dense systems, requires up to 50,000 tons of copper. This isn't just incremental growth; it's a novel, concentrated demand stream that is dragging the global balance sheet deep into the red. Even the more conservative ICSG forecast of a 150,000-ton deficit signals a market entering a structural shortage for the first time since 2009.

Supply, however, is struggling to keep pace. The sector is weighed down by prolonged mine closures and repeated downgrades. The Grasberg Block Cave remains shuttered until mid-2026, and Anglo American's recent guidance cut for its Chile operations underscores the difficulty of replacing lost output. This operational fragility is a critical vulnerability, as it means the market lacks the buffer of surplus supply that analysts once expected.

The stress is now evident in the physical midstream. While demand for refined copper is soaring, the supply of the raw material-copper concentrate-faces a different problem. Chinese smelters have added significant capacity, but they are finding it harder to secure the by-product-rich concentrate needed to feed it. This imbalance has driven treatment and refining charges (TC/RCs) to record lows, with the annual benchmark settling at $0 per tonne in January. Negative spot fees have been the norm since 2024. In essence, smelters are paying to take concentrate, a clear signal of a tight supply of the raw material that feeds the entire downstream chain.

The bottom line is that the long-term structural deficit narrative is gaining momentum. The AI-driven demand surge is real and powerful, while supply-side constraints are persistent and tangible. This sets up a market where the physical balance is under severe stress, even if near-term price moves are still influenced by inventory overhangs and cyclical demand shifts. The era of abundant copper may be over.

Catalysts and Watchpoints for 2026

The market's trajectory this year will be dictated by a handful of specific events that will test the competing surplus and deficit theses. The single most immediate catalyst is the timing of the US tariff announcement. Goldman Sachs' base case is for a 15% tariff to be announced in mid-2026. Any delay in that announcement, or in its eventual implementation, could dramatically prolong the current price weakness. The tariff uncertainty has already fueled a wave of speculative stockpiling, and its resolution is key to whether that artificial demand overhang persists or fades.

Closely tied to that is the health of Chinese demand, the world's largest consumer. The market has already seen a sharp contraction, with refined copper demand falling to -8% year-on-year in the fourth quarter of 2025. The key watchpoint is whether this decline stabilizes or worsens in the coming quarters. A recovery in Chinese consumption would be a major bullish signal, directly challenging the surplus forecasts and supporting the structural deficit narrative.

On the supply side, investors must monitor updates on mine production guidance and new project timelines. The narrative of a struggling supply chain is supported by prolonged closures like Grasberg and downgraded output from major operators. Any new guidance that confirms these constraints-or, conversely, any positive updates that suggest a faster-than-expected ramp-up-will be critical for validating or undermining the deficit forecasts.

Finally, the sheer scale of AI-driven demand remains a watchpoint in its own right. While J.P. Morgan's projection of a 330,000-metric-ton shortfall is extreme, the underlying assumption is that data center construction will proceed at a breakneck pace. Tracking the progress of these projects and the resulting copper consumption will provide real-time data on whether this novel demand stream is materializing as expected. The year's price action will likely be a tug-of-war between the near-term tariff overhang and Chinese demand, but the ultimate direction will hinge on which of these catalysts-structural deficit or cyclical surplus-proves more powerful.