- The G7 launched its Critical Minerals Resilience and Production Alliance on June 17, 2026, targeting a reduction in single-supplier dependence for rare earths and permanent magnets to below 60% by 2030, directly addressing a chokepoint in AI chip and data centre supply chains.
- China currently controls an estimated 70% of rare earth mining and close to 90% of processing capacity, according to reports, a concentration the alliance’s IEA-backed coordination platform and lithium and nickel pilot programmes are designed to counter.
- New mines take up to 30 years from discovery to full production in some cases, meaning the G7’s 2030 targets depend heavily on accelerating permitting and securing long-term purchase commitments from chipmakers and cloud infrastructure vendors, not just announcing investment pledges.
The G7 has a rare earth problem, and AI hardware sits at the centre of it. At their June 17, 2026 summit in Évian-les-Bains, France, the seven nations formally launched the Critical Minerals Resilience and Production Alliance, a direct response to the concentration of rare earth mining and processing that underpins everything from advanced semiconductors to data centre components. The timing matters: AI infrastructure demand is accelerating faster than the supply chains feeding it.
G7 Unveils Critical Minerals Alliance
The alliance sets a concrete goal: reduce reliance on any single supplier outside the G7 and its partners for rare earths and permanent magnets to below 60% by 2030, with a longer-term target of 50% as soon as possible. That framing points squarely at China, which reportedly mines around 70% of all rare earths and holds close to 90% of global processing capacity. That processing dominance is the harder problem. Mining can be relocated; refining infrastructure takes years and significant capital to build.
The alliance builds on the G7 Critical Minerals Action Plan from the June 2025 Kananaskis Summit in Canada, which set out ambitions to diversify mining, processing, manufacturing and recycling across allied nations. China’s past export restrictions on rare earth-related products and permanent magnets have already disrupted global industries, giving the 2026 alliance a concrete precedent to point to.
Bolstering AI and Chip Supply Resilience
Rare earths, lithium, nickel and cobalt are not abstract policy concerns, they are physical inputs to the GPUs, power systems and cooling infrastructure inside every major AI data centre. Concentrated supply chains for these materials create a direct risk to AI hardware production timelines, which is why supply chain resilience has become a strategic priority alongside chip design and fabrication capacity.
To coordinate the response, the G7 will establish a new platform for data sharing and joint decision-making among member states and partners. The International Energy Agency will contribute market analysis and early warnings on supply disruptions, shortages and price distortions. Initial pilot programmes focus on lithium and nickel, with plans to expand to five additional minerals annually, with rare earth elements the priority.
Financing Diversification and Investment Needs
Apple‘s July 2025 commitment to purchase $500 million in rare earth magnets from MP Materialstied to facilities in Texas and California, is the clearest model of what the G7 is trying to replicate at scale. Direct offtake agreements give miners and refiners the revenue certainty needed to justify capital expenditure. Without that, the economics of building new processing capacity outside China are difficult to make work.
Global demand for critical minerals is projected to increase significantly, though the specific trajectory depends heavily on clean energy and AI infrastructure build-out rates.
Overcoming Structural and Permitting Hurdles
The hardest constraint is time. In the United States, federal permitting for a new mine can take seven to 10 years under standard timelines, and some cases stretch up to 30 years from initial discovery to full production. Against a 2030 target, that arithmetic is brutal. Any mine that does not already have a permit in hand is unlikely to contribute meaningful supply within the alliance’s first planning cycle. Streamlining regulatory processes and reducing trade barriers among allied nations would help, but neither happens quickly.
The G7’s approach has also drawn criticism for its limited engagement with resource-rich developing countries. The risk, as some critics have framed it, is that the alliance functions as a consumer club, coordinating purchases and standards among wealthy importers while leaving producer nations without a meaningful seat at the table. A supply chain that bypasses China but still treats producing regions as pure extraction zones solves one geopolitical problem while creating another. For the alliance to hold, it will need to co-develop supply chain standards and support infrastructure, skills and regulatory capacity in mineral-producing regions, not just coordinate among importers.
The G7’s minerals push matters to AI hardware precisely because the bottleneck is not just fab capacity or chip design, it is the raw material layer beneath both. Whether the alliance moves fast enough to matter before the next round of export restrictions is the open question. For more coverage of AI chips and infrastructure, visit our AI Hardware section.
