AI and Commodities: The 10 Raw Materials Set to Surge
As the global race toward artificial intelligence (AI) supremacy accelerates, so too will the demand for the raw materials that power its infrastructure. From energy and server hardware to autonomous robotics and edge devices, the rise of AI over the next decade is expected to trigger one of the most significant resource expansions since the industrial revolution.
Top 10 Commodities Poised to Benefit
- Natural Gas: Data centres need reliable 24/7 baseload power. Analysts project AI-related data centres will drive an additional 3.3 billion cubic feet/day of global gas demand by 2030.
- Copper: Used in electrical wiring, transformers, and data infrastructure. Expected demand increase of 2–4 million tonnes by 2030 due to AI and electrification.
- Lithium: Essential for lithium-ion batteries powering AI servers, robots, and devices. Demand is expected to rise from ~180 kt (2023) to 500–850 kt by 2035.
- Nickel: High‑performance batteries require nickel‑rich cathodes. Global demand may rise from 3.35 Mt (2023) to 9.7 Mt by 2035.
- Cobalt: Offers battery thermal stability. Projections show demand could more than double to 530 kt by 2035.
- Rare Earth Elements (REEs): Elements like neodymium and dysprosium are vital for motors and cooling fans in AI systems and robotics.
- Silicon: Semiconductor-grade silicon demand will grow with every new AI chip and photonic processor. A 5–10 % increase is expected through 2035.
- Gallium: Crucial in compound semiconductors (GaN/GaAs) used in power and optical components. Demand could grow by 10–20 %.
- Uranium: AI data centres need constant power. Tech giants are investing in nuclear SMRs, with uranium demand expected to rise to ~95 kt/year by 2035.
- Aluminum: Used in server racks, transformers, and grid hardware. AI-related growth could add a further 5–10 % to global demand.
AI’s Fuel of Choice: Natural Gas
The chart below illustrates projected AI-driven natural gas demand growth, underscoring why reliable power remains a cornerstone of global AI infrastructure expansion:
By the Numbers: 2023 vs. 2035 Demand Outlook
| Commodity | 2023 Demand | 2035 (Low) | 2035 (Mid) | 2035 (High) |
|---|---|---|---|---|
| Nickel | 3.35 Mt | 6.5 Mt | 7.5 Mt | 9.7 Mt |
| Cobalt | 210 kt | 300 kt | 400 kt | 530 kt |
| Copper | 23 Mt | 30 Mt | 34 Mt | 38 Mt |
| Uranium | 67.5 kt U | 75 kt U | 85 kt U | 95 kt U |
| Lithium | 180 kt LCE | 500 kt | 650 kt | 850 kt |
What’s Driving This Surge?
- AI data centres may consume up to 12 % of U.S. power by 2030.
- Semiconductors require ultra-pure silicon and gallium compounds to support AI chips.
- Robotic and autonomous platforms will need durable, high-performance battery metals.
- Grid infrastructure build-outs depend on copper and aluminum expansion.
- SMRs and nuclear are being pursued by companies like Microsoft and Amazon to power AI clusters.
The Bigger Picture
AI is not just a software revolution—it’s a physical one. The materials listed above represent the foundational resources required to bring that transformation to life. Investors, policymakers, and strategic planners should be tracking these demand shifts closely as global competition intensifies around energy, infrastructure, and raw materials.
Sources: IEA, Goldman Sachs, Trafigura, Cobalt Institute, Sprott Asset Management, BloombergNEF, U.S. Geological Survey
