AI has become both a customer and a competitor of mining at the same time
AI data centers are driving up demand for copper, lithium and aluminum — and competing with mining projects for the same electricity both need to operate.
- The International Energy Agency projects that global data center electricity consumption should nearly double by 2030, reaching about 950 terawatt-hours — close to 3% of world demand.
- This growth is pushing up demand for copper, lithium, aluminum, uranium and rare earths, but data centers are also competing with miners for the same electricity, labor and transmission infrastructure.
- In Alberta (Canada), a $13 billion Meta AI data center led the grid operator to impose a 1,200-megawatt cap through 2028 — directly affecting lithium mining projects in the region.
- Industry executives already treat electricity as a strategic resource, on the same level as the critical minerals mining extracts.
The boom in data centers built to train and run artificial intelligence models has created a dynamic little discussed until now: these data centers depend on the same metals, the same electricity and, in many cases, the same regions that mining uses to operate. A recent survey combining data from the International Energy Agency, S&P Global and mining executives shows how this competition is already changing investment and energy-policy decisions across several jurisdictions.
The central point is that AI has come to occupy two opposite roles in mining's value chain at the same time — customer and competitor. As a customer, it drives demand: S&P Global projects global copper consumption will rise 50%, from 28 million tonnes in 2025 to 42 million in 2040, with the risk of a 10-million-tonne deficit if investment in new mines doesn't keep pace. Data center battery backup systems are shifting from 4-hour to 8-hour autonomy, which doubles lithium demand in those systems. Aluminum and uranium round out the list, driven by data hall infrastructure and power generation. As a competitor, AI competes for the same input every mine needs: firm, cheap electricity. In Alberta, Meta's 1-gigawatt data center near Edmonton sits 200 km from E3 Lithium's lithium project — and the province's grid operator has already imposed a 1,200-megawatt cap on new connections through 2028, a limit that weighs directly on both types of project. In British Columbia, BC Hydro reserved 400 megawatts of capacity for technology projects over two years, but exempted mining, forestry, manufacturing and LNG from that same cap — a regulatory choice that favors one sector over another in the queue for power.
For those planning mining projects, electricity has stopped being an operating cost line item and become a factor of strategic viability, as relevant as ore quality or metal price. Chris McCleave, technical director at Vale Base Metals, sums up the problem: electricity is now a strategic resource, on the same footing as the critical minerals mining itself extracts. This changes how projects should be evaluated — it's not enough to calculate reserves and grade, you need to map the fight for grid capacity in the region before committing capital. It also exposes a reputational tension: according to Chris Doornbos, CEO of E3 Lithium, data centers that arrive in a community without the same engagement care mining has historically had to show end up tarnishing the reputation of the entire heavy infrastructure sector by association.
What did we learn?
- Electricity has become a strategic input for mining, in the same category of relevance as mineral reserves and ore grade — and increasingly contested by other energy-intensive sectors.
- Energy regulators are already creating sector-differentiated rules (like Alberta's cap and BC Hydro's exemption), and that can determine which mining projects get off the ground first.
- AI demand for copper and lithium is real and measurable, but it comes with a risk rarely discussed: direct competition for the power grid, skilled labor and public support for large infrastructure projects.
Skills Radar
- Market Analysis★★★★★
- Project Planning★★★★★
- Risk Management★★★★★
Skills Developed
- Market Analysis
- Project Planning
- Risk Management
Upward trend
The projected growth of AI data centers through 2030 is structural, not cyclical — the fight over electricity and critical metals between mining and the tech sector is likely to intensify, not ease, in the coming years.
Who is this content useful for?
- Managers
- Executives
- Engineers
- Researchers
- Companies
To go deeper on this topic
Worth pursuing training in:
- Mineral and resource economics
- Mining project planning
- Energy and industrial infrastructure
- Institutional relations and permitting


