AI as a new driver of rare earth demand: supply risk lens for mine planners
Reviewed by Tom Sullivan
First reported on MINING.com
30 Second Briefing
Artificial intelligence is emerging as a third structural driver of rare earth demand alongside defence and electrification, with the five largest hyperscale data-centre operators planning about US$400 billion of 2025 capex and AI data centres forecast by the IEA to consume 3% of magnet rare earths by 2030. Sprott’s Steve Schoffstall links this to China’s control of about 94% of global permanent magnet production and recent Chinese export controls on military-related products. Western responses include US equity support and neodymium-praseodymium price guarantees for MP Materials, plus similar diversification moves in Australia and Japan.
Technical Brief
- Cooling systems in hyperscale data centres, consuming ~20% of site power, increasingly rely on rare earth magnets.
- Sprott notes rare earth magnets are now embedded in data-centre cooling, storage, semiconductors and comms hardware.
- Defence demand is underpinned by ~US$2.6 trillion annual global military spending and higher NATO targets.
- NATO members are moving defence budgets towards 5% of GDP, locking in magnet demand for guided weapons and aircraft.
- The US currently mines ~13% of global rare earths yet still imports about two‑thirds of its needs from China.
- China controls ~94% of permanent magnet production and has imposed export controls on military‑related rare earth products.
- Washington’s intervention includes equity in MP Materials, neodymium‑praseodymium price guarantees and loan‑backed processing plants.
- Average mine development from discovery to production is quoted at ~16 years, constraining rapid non‑Chinese supply growth.
- China has also restricted export of separation know‑how, adding a processing-technology bottleneck to Western refining plans.
- For project developers, long lead times plus state price guarantees signal a policy‑driven, not purely cyclical, rare earth market.
Our Take
Sprott Asset Management also features in our uranium coverage, including a 2025 piece on AI-driven power demand for hyperscale data centres, signalling that the firm is positioning AI not just as a rare earths theme but as a cross-commodity demand driver for both critical metals and nuclear fuel.
With China holding about 94% of global permanent magnet production and the US at 13% of mined rare earths, any AI-led uptick in magnet rare earths demand by 2030 heightens the strategic gap between mining capacity in the United States and downstream processing dominance in China.
The roughly 16-year average mine development timeline in the article’s data implies that projects targeting neodymium-praseodymium and other magnet rare earths now are unlikely to be fully responsive to AI data-centre demand peaks expected by 2030, favouring brownfield expansions and non-Chinese processing upgrades in the near term.
Prepared by collating external sources, AI-assisted tools, and Geomechanics.io’s proprietary mining database, then reviewed for technical accuracy & edited by our geotechnical team.
Related Articles
Related Industries & Products
Mining
Geotechnical software solutions for mining operations including CMRR analysis, hydrogeological testing, and data management.
CMRR-io
Streamline coal mine roof stability assessments with our cloud-based CMRR software featuring automated calculations, multi-scenario analysis, and collaborative workflows.
HYDROGEO-io
Comprehensive hydrogeological testing platform for managing, analysing, and reporting on packer tests, lugeon values, and hydraulic conductivity assessments.
GEODB-io
Centralised geotechnical data management solution for storing, accessing, and analysing all your site investigation and material testing data.