Why copper theft is spiking and what it says about the future

Next time you experience an internet outage, don’t blame the weather. Blame the thieves who took a big bite out of your internet line instead. According to Bloomberg, telecom companies are sitting on 800,000 tonnes of copper worth $7 billion at today’s prices in the form of old copper wiring. A surge in copper thefts is usually correlated with two things: a rise in copper prices and a rise in unemployment. We have seen a clear upward trend with the former and copper thefts could be a leading indicator of the latter.

There is a good chance we will be hearing more on cable-cutting thievery in the future. Copper traded around $3.89 per pound at the beginning of the year and recently crossed $5 per pound. So, what is going on with the price of copper?

Net zero targets come at a cost

First, it’s important to understand why copper is so central to our lives. Apart from silver, it’s the most efficient conductor of electricity. Copper is everywhere, from small appliances like toasters, to cars, which have an average of 29 kilograms of copper built into them. Electric cars need at least double the copper.

To meet ambitious net zero targets, it’s not just about electrifying means of transport. It also means generating and sourcing clean energy from far-flung wind farms and solar parks that need millions more feet of copper wiring to connect greater distances. According to S&P Global, annual copper demand is expected to double to 50 million tonnes by 2035 to achieve net zero targets.

Power-hungry AI search

Your Google search consumes around 0.0003 kilowatt-hours of energy. Ask the same question on, say, ChatGPT4 and energy consumption jumps to 0.01 kilowatt-hours. That’s around 15 times more energy consumption. Extrapolate this to the exponential growth in AI searches we see on the horizon (ChatGPT had 100 million monthly active users at the end of 2023) and you get the scale of energy required to sustain this trend. According to Trafigura, copper demand related to AI and data centres could add up to a million tonnes by 2030. That’s on top of the expected four to five million tonne deficit gap expected for reasons other than AI.

Structural deficits

While demand could explode on the back of AI and ambitious net zero goals, supply is not expected to keep pace for two reasons. First, the 10 biggest copper mines are around 95 years old on average. With every passing year, it gets more expensive to go deeper into the ground to extract lower grades of ore. Most of the big mines are located in Chile and Peru, politically volatile jurisdictions with recurring water shortages. That leaves existing mines producing a dwindling supply of expensive copper.

When it comes to new mines, there is not much incentive to put dollars to work on known deposits. Copper prices would need to be much higher to incentivize mining companies to invest billions of dollars. With politics, red tape and tougher environmental regulations, it takes close to a decade to fully operationalize a mine. This explains the flurry of activity in recent months (Glencore & Teck, Newmont & NewCrest, BHP & Anglo) to buy and consolidate existing mining assets. However, without exploration for new assets, it’s unlikely this will make a dent in the current demand-supply equation. The last copper mega deposit was found in Southern Mongolia (Oyu Tolgoi mine) 20 years ago. After pouring $7 billion into it, it has yet to reach peak production.

One of our holdings is Capstone Copper (CS CN), a mid-tier copper producer with mines in Chile, Mexico and Arizona. Capstone’s growth is driven by its assets in Chile, which make up two thirds of its net asset value. As a mid-tier producer, Capstone is not blessed with scale or the most unique grade of ore. However, what we like about the company is its ability to find and manage high IRR assets stewarded by a CEO with over 30 years of experience in the mining space.

In 2023, the company produced 164 kilotonnes of copper; however, the ramp up of its two Chilean mines is projected to bring production to 350 kilotonnes by 2028. This production increase means cash costs could come down from $2.88 per pound today to $1.50 per pound in the future. Should copper prices remain elevated, this would make for a very profitable operation.

Twenty years ago, copper saw a five-time increase in price driven by underinvestment in the 90s and China’s rapid industrialization. Could the simultaneous confluence of AI, the green energy transition and supply shortages make history repeat itself?