Here's Why I Like This Copper ETF More Than Ever

Copper prices rose 48% over the past year to roughly $6.89 per pound in early September, near an all-time high, as supply discoveries have collapsed while AI data-center demand accelerates. No major copper deposits were discovered in 2025, and new projects take an average 17.5 years to reach production; AI data centers can require up to 50,000 tons of copper each versus 5,000-15,000 tons for conventional facilities. The Global X Copper Miners ETF (COPX), with about $8.4 billion in assets, has gained roughly 24% year-to-date and 64% over the past year, though the article notes potential near-term volatility from changing AI-investment sentiment.

Analysis

The investable question is no longer whether copper is structurally scarce, but whether spot pricing is sufficiently above miners' incentive price to convert scarcity into incremental free cash flow. FCX offers the cleanest liquid beta, but its earnings sensitivity is partly offset by operating/geopolitical exposure at Grasberg; SCCO has higher copper torque but carries concentrated Peru/Mexico jurisdiction and a more demanding valuation. HBM is a higher-beta alternative where a sustained copper-price regime could accelerate deleveraging and rerating, although its weaker balance sheet makes it materially more vulnerable to a copper drawdown.

The second-order beneficiary is not necessarily the miner ETF: higher copper costs compress electrical-equipment, grid, cable, renewables and EV supply-chain margins before project repricing catches up. AI infrastructure is comparatively inelastic at the hyperscaler level, so the near-term burden should fall on EPCs and hardware vendors with fixed-price contracts rather than NVDA; over 6-18 months, persistently high copper raises data-center capex intensity and could modestly slow marginal project returns. Supply constraints are also prone to political repricing: a permitting or fiscal-policy shift in Chile, Peru, Panama, or the DRC can move the forward curve more than exploration headlines.

Consensus appears to be extrapolating a structural narrative from a sharply risen spot price. Over the next 1-3 months, Chinese property/import data, LME/SHFE inventory trends, and the dollar will dominate the marginal copper price more than long-dated mine-development constraints. The bullish thesis is falsified tactically by a sustained inventory rebuild alongside weaker treatment charges and FCX/SCCO declining to reaffirm volume and unit-cost guidance; structurally, it is challenged if AI capex converts to less copper-intensive architectures or major brownfield expansions arrive ahead of schedule.

Key Decisions for Investors

• Prefer a 6-12 month long FCX over COPX for liquid copper exposure; enter on a 8-12% copper/FCX pullback rather than chase momentum. Target a 15-25% equity upside under sustained elevated copper pricing; exit if FCX cuts annual production guidance or copper breaks below its 200-day moving average on rising exchange inventories.
• Use a barbell: long FCX and a smaller HBM position, sized at roughly 2:1, rather than concentrated SCCO. HBM provides higher FCF/deleveraging convexity if copper remains firm, but cap position size given financing and execution risk; reassess after each quarterly net-debt and cost update.
• For a relative-value expression over 3-6 months, long FCX / short a diversified industrial-input basket such as XLI only if copper inventories continue drawing. The expected mechanism is widening raw-material cost pressure versus miners' operating leverage; abandon the pair if industrial pricing power offsets input inflation or copper inventories reverse higher.
• Do not add AI semiconductor exposure solely on the copper narrative. Monitor hyperscaler capex guidance and copper intensity per megawatt; any evidence that copper-driven capex inflation is constraining build schedules would be a negative second-order signal for marginal data-center equipment demand, not a direct reason to sell NVDA.