For decades, the commodities market followed a predictable pattern: demand rose as emerging markets became more industrialised, and technological advances slowly lowered the cost of extraction. But as we get closer to 2026, something new has broken this cycle. We are seeing the AI Power Paradox, which is a structural change in which AI is both the main cause of commodity scarcity and the best way to fix it.
The AI Power Paradox isn’t something that needs to be “solved.” It’s something that needs to be managed.
Producers have a clear job: they need to turn into a tech company that also mines. Investors are no longer interested in “who has the most ore” but rather “who has the best algorithms to get it out.” This isn’t just “tech meeting dirt.” It changes the basic equation of supply and demand around the world.
The Hunger: AI As A Source of Goods
AI is in the “cloud,” but its roots are deep in the ground. The huge growth of data centres needed to train Large Language Models (LLMs) has made the market hungry for materials that it can’t keep up with.
The Copper Shortage
In the digital age, copper is like the nervous system. AI infrastructure uses a lot of copper, from the high-voltage connections that link data centres to the grid to the thick cables that run through server racks. According to current estimates, AI data centres could need up to 1 million tonnes more copper each year by 2030.
The Bridge Of Natural Gas
Even though there is a push for renewable energy, AI workloads need baseload power that wind and solar can’t provide on their own yet. This has caused a huge increase in the need for natural gas in the U.S. and Europe. We are seeing a “lock-in” effect where tech giants, who used to be the biggest supporters of green energy, are now signing long-term contracts for gas-fired power to make sure their “AI factories” never go dark.
More from Artificial Intelligence
The Efficiency: AI As The Great Extractor
The paradox is that AI is running out of resources, but it is also the only technology that can find more. The mining industry is going through a “digital renaissance” right now, using AI to deal with the problems of lower ore grades and deposits that are getting harder to reach.
Digital Twins And The Speed Of The Supply Chain
The Digital Twin, which is a real-time virtual copy of a physical supply chain or mine, is one of the most important tools in 2025. By combining AI with IoT sensors, operators can run thousands of “what-if” scenarios, such as delays in logistics due to bad weather or broken equipment.
- Maintenance that is based on predictions: AI can now tell when a $5 million haul truck will break down weeks before it happens, which can cut down on downtime by up to 30%
- Precision Mining: AI algorithms look at hyperspectral satellite images to find geological anomalies that human geologists might miss. This cuts down on exploration time by almost 40%
The Jevons Paradox Warning
As thought leaders, we need to be aware of the Jevons Paradox, which is a historical trap. It says that when a technology makes a resource easier to use, the total use of that resource goes up because the lower cost makes people want it more.
Will we just build more data centres as AI makes mining easier and cheaper to get copper? This will put even more strain on the planet’s resources. This “rebound effect” is the biggest problem that commodity strategists will face in the next ten years.
More from Artificial Intelligence
The “Synthetic Supply” Plan
The industry’s main goal is to make “Synthetic Supply.” This isn’t about making fake copper; it’s about using AI to make things more efficient so that we can meet the needs of the digital economy with what we can physically dig out of the ground.
Example: Autonomous Fleets
AI-powered autonomous fleets are now everywhere at big sites like BHP’s Escondida and Rio Tinto’s Gudai-Darri. These systems do more than just “drive themselves.” They make the most of every gram of fuel and every minute of travel time. This level of optimisation makes “extra” supply by making sure that no energy or material is wasted during the extraction process.
