Over the past two decades, the world grew accustomed to the "asset-light" economic miracle represented by companies like Google, Meta, and SaaS models. However, with escalating geopolitical conflicts, supply chain restructuring, and the explosion of the AI computing revolution, the world is now facing an unprecedented tsunami of demand for heavy assets and commodities.
Dan Dreyfus, a veteran Wall Street commodities expert and partner at Fortnite Capital, has clearly identified that the next core bottleneck for AI is not limited to memory like HBM, but the most fundamental "king of metals" – copper. He projects that humanity's demand for copper over the next 18 years will equal the total of the past 10,000 years, while current power grids and critical mineral supplies are on the brink of extreme fragility.
The Era of Light Assets is Over: A Reckoning for Heavy Infrastructure
Dreyfus explains that from the early 2000s until a few years ago, the US and the world experienced an "asset-light" economic miracle. Giants like Google, Meta, and Apple created trillions in market value with minimal capital investment. Yet, while doubling down on this asset-light mentality, Western nations were simultaneously dismantling critical infrastructure and offshoring it.
This fragile, non-resilient supply chain began to backfire with successive shocks from the COVID-19 pandemic, the Russia-Ukraine war, tariffs, and Middle East tensions, leading to soaring and persistent inflation.
The world is now at a critical inflection point: reshoring of manufacturing, reindustrialization, and an "AI technological revolution" that is far more infrastructure-intensive than previous computing generations are all occurring simultaneously. This is creating an incredibly intense demand shock for critical minerals and commodities.
Currently, multiple mega capital cycles are running concurrently: the aerospace and space economy, with Boeing and Airbus facing a $1 trillion order backlog over the next decade and competition for raw materials from the space economy; power grid modernization, with aging grids in the West struggling even before the AI electricity tsunami; and power generation & data centers, where infrastructure and commodity spending is already massive.
The King of Metals Faces a Supercycle: AI Data Centers to Consume 750,000 Tons of Copper Annually
In this infrastructure renaissance, copper, as the "king of metals," faces an unprecedented supply-demand mismatch. Both clean energy and AI exponentially increase copper consumption: solar power requires 5 times more copper per megawatt than traditional gas turbines; wind power requires 7 times more; and a 1-gigawatt (GW) AI data center factory requires 50,000 tons of copper. With the industry expected to build 15 GW of such facilities annually, data centers alone will add 750,000 tons of new copper demand per year. Electric vehicles consume 5-6 times more copper than internal combustion engine cars, and modern military conflicts consume vast, non-recyclable amounts of copper in munitions.
Dreyfus emphasizes a staggering comparison: over the past 10,000 years, humanity mined 700 million tons of copper. Current annual global copper demand is 30 million tons. Even ignoring the extra demand from data centers and green energy, and merely following traditional GDP growth, the world will need 700 million tons of copper over the next 18 years – matching the total mined in the past ten millennia.
Meeting this demand would require five new world-class, Tier-1 mines to come online every year. However, with aging mines declining in ore grade and new mines taking 7-12 years to build, very few such mines are slated for production before 2030. While market focus is on memory, HBM, and NAND, the next looming critical bottleneck is undoubtedly copper. Dreyfus states plainly, "I think it's very easy for copper to double from here."
Harsh Realities of the Energy Pool: Solar and Nuclear Can't Provide Instant Solutions
While solar and nuclear power are often seen as solutions to the massive power shortfall, they face significant raw material and physical constraints. Solar power has a capacity factor of only about 20%. To power a 1 GW data center entirely with solar would require 5 GW of solar capacity, needing 35,000 acres of land – larger than the city of San Francisco. Furthermore, a silver crisis looms: solar panel manufacturing requires vast amounts of silver. Global annual consumption is 1.2 billion ounces against a supply of only 1 billion, creating a 200 million ounce annual deficit. With only 600 million ounces of above-ground inventory left, the solar supply chain faces a potential break within three years without a solution.
For nuclear power, while natural gas and uranium reserves are ample, the US currently cannot manufacture key heavy industrial components like containment pressure vessels domestically, creating severe supply chain dependencies.
The Macro Backdrop: Dollar Devaluation and "Hard Assets" as the Only Haven
Beyond intrinsic supply-demand shocks, macro monetary policy is "adding fuel to the fire" for a commodity surge. Dreyfus notes that since the pandemic, the value of fiat currency has been massively diluted. US government debt stands at $40 trillion, growing by $2.5 trillion annually, while the present value of future social liabilities like Medicare and Social Security is $100 trillion, also growing by $2.5 trillion yearly. Contrasted with annual US tax revenues of only $5.5 trillion, the next economic downturn will likely force the Federal Reserve to "print giga dollars." In a similar 1970s environment, currency lost 70% of its purchasing power. In such conditions, commodities, hard assets, and critical infrastructure are the only solutions to preserve purchasing power.
The Return of Labor to the Throne: Skilled Blue-Collar Workers Become the Scarcest Resource
In this revival of heavy industry and infrastructure, skilled craft labor has replaced white-collar workers as the biggest bottleneck in the entire US and global supply chain. Over the past decade-plus, societal trends steered youth towards liberal arts and asset-light industries, causing a severe generational gap in skilled trades. The situation is now completely reversed: demand for skilled blue-collar workers in infrastructure is nearly limitless. Top graduates from vocational technical colleges can command starting salaries of $150,000 right out of high school. Ironically, the same blue-collar workers from the heartland who were previously displaced are now commanding premium wages for work that cannot be replaced by AI in the short term, while earlier-stage, lower-level white-collar jobs face AI optimization.
Comments