91探花

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鈥檛 something that needs to be 鈥渟olved.鈥� It鈥檚 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 鈥渨ho has the most ore鈥� but rather 鈥渨ho has the best algorithms to get it out.鈥� This isn鈥檛 just 鈥渢ech meeting dirt.鈥� It changes the basic equation of supply and demand around the world.

听

The Hunger: AI As A Source of Goods

听

AI is in the 鈥渃loud,鈥� 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鈥檛 keep up with.

听

The Copper Shortage

听

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.

听

The Bridge Of Natural Gas

听

Even though there is a push for renewable energy, AI workloads need baseload power that wind and solar can鈥檛 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 鈥渓ock-in鈥� 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 鈥淎I factories鈥� never go dark.

The Efficiency: AI As The Great Extractor

听

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 鈥渄igital renaissance鈥� right now, using AI to deal with the problems of lower ore grades and deposits that are getting harder to reach.

听

Digital Twins And The Speed Of The Supply Chain

听

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 鈥渨hat-if鈥� 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%

听

The Jevons Paradox Warning

听

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鈥檚 resources. This 鈥渞ebound effect鈥� is the biggest problem that will face in the next ten years.

The 鈥淪ynthetic Supply鈥� Plan

听

The industry鈥檚 main goal is to make 鈥淪ynthetic Supply.鈥� This isn鈥檛 about making fake copper; it鈥檚 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.

听

Example: Autonomous Fleets

听

AI-powered autonomous fleets are now everywhere at big sites like BHP鈥檚 Escondida and Rio Tinto鈥檚 Gudai-Darri. These systems do more than just 鈥渄rive themselves.鈥� They make the most of every gram of fuel and every minute of travel time. This level of optimisation makes 鈥渆xtra鈥� supply by making sure that no energy or material is wasted during the extraction process.