3 Trends That Will Push Copper Higher in 2028

Copper demand could rise 50% by 2040, with AI and data-center demand projected to triple; a hyperscale facility can require up to 50,000 metric tons of copper. U.S. electricity demand is forecast to grow 2.5% annually through 2035—five times the prior decade's rate—driving grid and renewable-infrastructure investment, while 44% of copper output already serves power applications. Declining ore grades in Chile and Peru and 16-18-year greenfield mine development timelines could create widening deficits from 2028, although brownfield expansions, recycling and weaker Chinese construction may provide temporary relief in 2027.

Analysis

The investable expression is not a blanket long in copper miners. A rising copper tape initially rewards high-beta producers, but sustained deficits disproportionately favor low-cost, long-life assets with expansion optionality and politically stable jurisdictions. Freeport-McMoRan (FCX) has the cleanest U.S.-listed torque; Lundin Mining (LUNMF) and Southern Copper (SCCO) offer asset-level leverage, while COPX dilutes this with meaningful exposure to higher-cost and jurisdictionally riskier operators. Grid buildout also creates a less cyclical second derivative for Prysmian (PRYMY), Nexans (NEXNY), Eaton (ETN), and Quanta Services (PWR), which can monetize copper intensity through equipment and installation margins rather than solely metal-price beta.

Near-term, copper is vulnerable to a China construction slowdown, a stronger dollar, and recycled-supply elasticity; these can overwhelm a multi-year scarcity narrative for several quarters. The more important 1-3 month catalyst is whether hyperscaler capex converts into signed utility interconnection, transformer, and transmission orders rather than remaining announced spend. Over 6-18 months, permitting delays and power shortages may constrain data-center commissioning, reducing the timing benefit to copper while increasing value capture for ETN, PWR, and electrical-equipment bottlenecks.

Consensus likely overstates direct data-center copper demand while understating the grid bottleneck and miner cost inflation. Lower grades raise energy, water, labor, and capex intensity, so copper-price upside does not translate one-for-one to miner FCF; producers with escalating unit costs can underperform the metal. Conversely, if copper remains range-bound, electrical infrastructure vendors may still rerate on backlog visibility and pricing power.

The thesis is falsified by a sustained deterioration in utility capex/backlog, material downward revisions to hyperscaler capital budgets, or a copper market surplus driven by Chinese demand weakness and faster-than-expected brownfield supply. For FCX specifically, watch unit-cost guidance and realized copper pricing at earnings; for the equipment complex, watch book-to-bill and backlog conversion rather than headline AI spending.

Key Decisions for Investors

• Initiate a 6-12 month pair: long ETN and PWR / short COPX in equal dollar risk. This captures grid bottlenecks with less Chinese-demand and mine-jurisdiction exposure; reassess if ETN/PWR backlog growth falls below management guidance or copper breaks materially higher on verified supply disruption.
• Build FCX opportunistically only on copper-price pullbacks or post-earnings cost-driven weakness, rather than chase a thematic rally. Target a 12-18 month position sized for high commodity beta; exit if FCX cuts volume guidance, raises unit-cost guidance materially, or copper inventories rise alongside weakening Chinese import data.
• For direct metal exposure, prefer staged long HG copper futures or COPX only after confirmation that the forward curve tightens and exchange inventories decline for several weeks. A 2027 supply response and recycling elasticity make immediate full-size exposure unattractive; use defined-risk call spreads rather than outright calls if implied volatility is elevated.
• Maintain a watchlist alert on MSFT, AMZN, GOOG, and META: a capex reduction alone is insufficiently bearish for copper, but a combination of lower capex and delayed data-center energization would be a signal to cut copper beta and add to the ETN/PWR relative-value leg.