AI development's other challenge besides power: minerals
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Aïda Amer/Axios
Everyone knows getting power for data centers is a challenge for U.S. AI development — but don't sleep on minerals, new and recent analyses warn.
Why it matters: A new Barclays report finds that obtaining critical minerals and attracting elite talent are the "new battlegrounds in the fight for AI supremacy."
State of play: On the materials front, over 60% of the most critical raw minerals for AI hardware come from a few developing and emerging countries, it finds.
- Materials like copper, cobalt, aluminum, lithium and nickel are needed for wiring and other components, battery materials, and storage infrastructure.
- Rare earths (not that rare, I know) are essential for semiconductors, Barclays notes. China recently imposed new export controls.
Threat level: In a vulnerability akin to risks for clean energy supply chains, China dominates the processing of many minerals produced elsewhere.
- For instance, it provides nearly 80% of processed cobalt and over 90% of refined rare earths needed for magnets.
What we're watching: The report provides various recommendations.
- It notes that mineral-rich nations like Chile and Congo have opportunities to expand their processing and boost geopolitical leverage.
- "They could secure strategic advantages by forming trade relationships and alliances with tech-heavy economies such as the US, the EU, and Japan," it states.
- Barclays also sees opportunities to boost e-waste recycling.
Catch up quick: "There is a significant overlap between the minerals needed for building new data centres and those that are critical to energy technologies," the International Energy Agency said in a report last month on the energy-AI nexus.
- IEA name-checks various material needs, like silicon for processors and memory components, aluminum for server racks, and gallium for power converters, to name just a few.
The bottom line: "Geopolitical conflicts, trade wars and natural disasters can turn dependencies into vulnerabilities," Barclays analysts write.
- "While high-tech chips and components often steal the spotlight, it is the critical mineral dependencies in the AI supply chain that are the true linchpin."
