Digital infrastructure is growing faster than the mining industry can supply it. According to Wood Mackenzie, more than 250 gigawatts of data center projects have been announced worldwide, three times as many as a year earlier. Yet the power grids and generation without which these facilities cannot operate will require many times more metal than the server halls themselves.
High commodity prices were supposed to spur the launch of new mines, but this is not happening. Wood Mackenzie Vice President Lindsay Grant explains the gap simply: demand is accelerating while delivery of resources is becoming harder. Permitting is dragging on, projects are getting more expensive, and geopolitics adds uncertainty. As a result, producers are not keeping pace with the ambitions of governments seeking to secure critical raw materials.
This is most visible in copper. Prices have reached $14,759 per tonne, but a significant part of the rise is explained not by the supply-demand balance but by a trade skew: the US has built up stocks sufficient for more than 170 days, while the rest of the world has only 48. As a result, a refined metal shortfall of about 450,000 tonnes is forming outside America. Chile will cut output by 5.2% in 2026, but in 2027 global output will grow by more than 5%. To maintain the pace, the industry needs to bring online roughly 850,000 tonnes of new capacity every year for ten years.
Aluminum took a hit from outside. Attacks damaged two plants in the Middle East, depriving the market of almost 2% of global production. A shortfall of 1 million tonnes is expected in 2026. Producers have responded: China has loaded its capacity to 99%, and Indonesia is raising output from 1.3 million tonnes to almost 3 million by 2030. By 2027, the market is forecast to return to surplus.
Nickel is emerging from a prolonged glut. In 2026, global refined metal production will fall by 2%, the first decline since 2015. The market has become stratified: Indonesian quotas, ore depletion, and droughts are squeezing Class II supply, where a deficit is expected from 2027. Chinese metallurgists are compensating with scrap and Class I cathodes. The Class I and chemicals segments are still in surplus. By 2030, the price could rise to $19,000 per tonne.
The battery segment is changing under the influence of lithium iron phosphate chemistry. LFP is forecast to dominate the mass-market segment through 2030. In parallel, CATL has signed a three-year contract for 60 gigawatt-hours of sodium-ion modules, with price parity with LFP expected by 2035. Solid-state batteries could enter series production around 2030.
Lithium carbonate has rebounded 130% from its 2025 lows, but raw material accounts for only 2–3% of the cost of a typical 100-megawatt storage system. The 83% rise in lithium prices in the first quarter of 2026 increased the price of a finished system by about 10%.
The main constraint is money. Over ten years, the lithium industry has raised $60 billion and increased chemical output tenfold. Another $150–250 billion will be needed through the 2050s. Chinese companies build capacity more efficiently than Western ones, which makes it harder to create independent supply chains outside China.
Energy is becoming a barrier in its own right. In Japan, a data center is built faster than grids are upgraded: less than five years versus more than ten. Developers are increasingly considering microgrids and standalone generation, although these cost two to three times more than connecting to the main grid.
Supply security, energy availability, and geopolitics now weigh as much as cost. The mining industry will have to either speed up project launches and simplify permitting, or accept that the digital boom will run into a physical shortage of metal.
Source: Wood Mackenzie
