AI and Copper

Popular business media companies like Fast Company, Business Insider, Reuters, Wall Street Journal, and Barrons all ran similar stories in the last few weeks.

The topic was an odd marriage of innovative technology and basic materials. 

The story covered how artificial intelligence (AI) would have a material impact on one of the metals that underpins our modern society: copper.
The copper price continues to climb as more investors do the math and work out the problem. We are going to need more copper than we currently produce. And that’s moving the price higher, as you can see here:

The latest trigger that led to several essays on Monday April 8th was a statement by giant commodity trading company Trafigura at the Financial Times Global Commodities Summit in Lausanne, Switzerland. Trafigura’s Chief Economist Saad Rahim said:

Copper demand linked to artificial intelligence and data centres could add up to one million metric tons by 2030 and exacerbate supply deficits towards the end of the decade.

That’s on top of a forecast four to five million metric ton annual supply gap by 2030. According to Rahim, this isn’t being factored into supply/demand models right now. 

The real driver for copper consumption is data center infrastructure. Giant tech firm Nvidia sparked some of this concern in March 2024. The company announced that it would switch from fiber optic cables to copper in AI data centers. The company said that moving to copper would help it cut power usage at its data center.

Power demand is at the center of this emerging story. According to the International Copper Association, data centers will account for 67% of copper demand in building by 2030. That’s up from just 37% in 2018. And they attribute that increase to the number of large and “hyperscale” data centers being built. 

For example, Vantage Data Centers, which builds and operates data centers globally, saw massive growth in 2023. The company broke ground on seven new hyperscale data centers. Bloomberg Intelligence expects hyperscale cloud services and generative AI to be a $1.3 trillion business by 2032. That’s driving massive construction of data centers around the world. 

These data centers are massive. For example, Vantage’s Quincy Washington campus has three facilities and cost more than $1 billion. Its Phoenix Arizona is still under construction. It’s a $1.5 billion facility scheduled to come online any day now. 

These are massive facilities that consume huge amounts of electricity…and therefore copper. This is the reality of our “New Energy” investment thesis. All this technology will be underpinned by simple, basic materials like copper. 

For the Good,

The Mangrove Investor Team

Numbers You Need to Know

20%

Over 20% of technology companies surveyed are aggressively pursuing integration of AI across a wide variety of technology products and business workflows. (CompTIA)

1 in 4

More than 1 in 4 dollars invested in American startups this year has gone to an artificial intelligence-related company, Crunchbase data shows. (Crunchbase)

85 to 134 terawatt hours

By 2027 the AI sector could consume between 85 to 134 terawatt hours each year. That’s about the same as the annual energy demand of an entire country.(The Verge)

What’s New in Sustainable Investing

Investor Trust in Sustainability Data : Key Takeaways From Global Study

Investors are increasingly incorporating sustainability factors into investment decisions and are seeking an opportunity, with an estimated US$43 trillion in global economic growth projected between 2021–2070, if the world economy transforms to achieve net-zero emissions. (Impakter)

Intel Guns for Nvidia With Gaudi 3 AI Chip

While the artificial intelligence (AI) accelerator market is dominated by Nvidia, Intel has been trying to gain ground. The company’s Gaudi line of AI chips historically couldn’t match Nvidia on performance, but Intel pitched them as a cost-effective alternative to Nvidia’s pricey graphics processing units (GPUs).(Yahoo Finance)

Video Of The Week