ANTHONY DE LA COUR: Green is the new gold

In a volatile year for commodities, copper emerged as one of the more compelling stories. Underpinned by a supercycle thesis, copper was fuelled by low inventories, a supply shortage of mined concentrate and the unexpected rise of the generative AI (GenAI) narrative.

The outcome was a largely speculation-led spike in the copper price to a new high in inflation-adjusted real terms. Though this has pulled back, the long-term investment case remains compelling.

A commodity supercycle occurs when long-term secular growth in demand exceeds supply, pushing prices above levels suggested by fundamental production costs. In copper’s case, the supercycle began in 2020, driven by the realisation that decarbonisation efforts would rely heavily on the carbon abatement strategies of electrification, renewable power and the rollout of supporting infrastructure.

Practically, achieving carbon net zero by 2050 involves the large-scale adoption of electric drivetrain vehicles (EVs), the expansion of renewable energy sources such as solar and wind, and the implementation of energy storage systems, all of which require substantial amounts of copper. Battery EVs use about five and a half times more copper than traditional internal combustion engines, while plug-in hybrid EVs use about four times as much. The same trend is evident in renewable energy: wind and solar power are nearly three times more copper-intensive than traditional energy sources.

Recent setbacks in the energy transition story notwithstanding, the bullish thesis is on track. The International Energy Agency predicts that by 2050 copper demand from electric drivetrains will grow from 2% of demand in 2023 to 13% in its net-zero scenario. EV sales increased by 35% in 2023. In 2010, renewable power accounted for 19.5% of global electricity production. By 2023 this had increased to 30% and Goldman Sachs estimates that, to meet targets, it would need to increase to 75% by 2050.

In 2024 a demand narrative that was dominated by energy transition pivoted for the first time to a new source: the huge “hyperscaler” data-centre warehouses that house the servers and processing units providing GenAI’s computational resources. GenAI’s power consumption is impressive. Goldman Sachs estimates that a ChatGPT query uses 10 times the power of a Google search, while a Carnegie Mellon study estimates that generating a single image using GenAI requires as much energy as half a smartphone charge.

While there is still considerable uncertainty around the potential effect of GenAI on copper demand, the forecasts are not inconsequential. For instance, Morgan Stanley estimates an additional 640kt of copper demand by 2027. Though this incremental demand is still lower than that from the energy transition, it represents a meaningful addition to the overall copper demand picture.

The supply response challenge

Despite ostensibly positive medium- and long-term demand signalling, the supply response to satisfy this has been underwhelming. Rising environmental, social, and governance (ESG) concerns and regulatory hurdles have increased the complexity and cost of successfully progressing investment decisions. Increased industry capital discipline and a focus on returns have worsened this. Factor in declining grades, an ageing asset base and long development times, and this is an industry potentially unprepared for the future.

The ability to develop growth volumes starts with the pool of potential deposits and the time taken to bring these to market. S&P Global sees a declining trend in copper discoveries, with those in the last decade accounting for only 14 of the 239 deposits discovered since 1990, and just four discoveries in the past five years, a result of the industry’s continued focus on brownfield assets at the expense of exploration.

While fewer major discoveries are being made, the time taken to develop them is also increasing. S&P Global notes that the average lead time, from discovery to production, of new mines has increased from 12.7 years to 17.9 years over the past two decades.

Despite this, the industry appears to have partially deferred long-term investment decisions. During the 2000s supercycle, the copper price rally incentivised a significant investment in new copper projects, and when prices subsequently crashed so did approvals. This saw a step change lower in the volume of mine capacity brought online from 2015. Though prices recovered in 2021, project approvals have yet to follow as companies prioritise capital discipline and shareholder returns.

In commodities, it is commonly understood that the cure for high prices is ... high prices. What differs this time (at least to date) is that supply has not yet responded to the current higher copper prices and, with lead times increasing, the window for new approvals to meaningfully bridge this gap is closing. This at a time when persistent deficits over recent years have reduced what inventory cushion the industry once had.

To satisfy future copper demand, prices are likely to need to remain high. The incremental tonnes required to satisfy this demand must come from brownfield operations that are higher cost or greenfield operations in riskier jurisdictions that are costlier to build than current operations. The market will need high prices now to have any hope of stimulating sufficient supply required for most base demand projections.

The copper price is therefore likely to continue to trade above the fundamental support that production costs would suggest.

• De la Cour is an investment analyst at Matrix Fund Managers.