show index hide index
As artificial intelligence continues to evolve at breakneck speed, a shadow hangs over this technological revolution: the metals crisis. Resources essential to the manufacture of critical components such as chips and servers are running out, and their supply is already affected by geopolitical tensions and fragile infrastructure. The question then arises: is AI, a symbol of a promising future, at risk of hitting a resource shortage wall?As artificial intelligence (AI) continues to transform the world, a major challenge looms on the horizon: the shortage of critical metals needed for its development. This phenomenon, linked to a plummeting supply and exponentially increasing demand, is jeopardizing long-awaited technological advances. The question arises: is AI about to stumble in the face of this metals crisis? A Hardware Infrastructure Under Pressure The rise of AI doesn’t rely solely on sophisticated algorithms or ever-increasing volumes of data. Indeed, the success of these technologies also depends on a complex hardware infrastructure. The latter requires so-called « strategic » or « critical » metals, without which innovations cannot see the light of day. For example, essential resources such as copper, aluminum, and cobalt are essential for manufacturing graphics processing units (GPUs) and the processors at the heart of AI. Metals in the SpotlightAmong the most sought-after metals, copper and aluminum play a central role. Used in crucial components such as cabling and heat dissipation systems, their excellent conductivity ensures optimal operation of data centers. Alongside them, materials such as tin, silver, and gold are also essential to ensure the
conductivity and resistance
to extreme conditions.
In addition, other metals such as tungsten and cobalt play a key role in the manufacture of electronic circuits, helping to meet the growing demands for miniaturization, essential for the development of AI technologies. Geopolitical tensions at the heart of the issue The production of these critical resources is highly geographically concentrated, highlighting security of supply issues. Indeed, nearly 70% of the world’s cobalt comes from the Democratic Republic of Congo, a country plagued by political instability. This situation makes supply extremely vulnerable to disruptions. Meanwhile, China holds a virtual monopoly on the refining of rare earths, controlling more than 80% of global production. This strategic power exacerbates tensions between countries and further complicates access to these resources for other nations, particularly due to measures such as taxes or export restrictions.Exploding Demand The forecasts are alarming. According to the International Energy Agency (IEA), global demand for copper linked to AI could grow by 40% by 2040, while that for rare earths could increase by 80%. As for lithium, its consumption is expected to explode, with an estimated increase of 700%. This increased competition is partly explained by the growth of data centers, which consume enormous quantities of metals. Raw material needs are colliding with complex market dynamics, exacerbated by the growing enthusiasm for other sectors such as renewable energy and electromobility. This increases the risks of shortages and price spikes, already evident in the AI industry. Adapting AI to the risks of shortages Faced with this challenge, avenues are emerging to mitigate the impacts of the critical metal shortage. Recycling and the circular economy appear to be promising solutions. By recovering the metals contained in end-of-life equipment, it would be possible to cover a significant portion of future demand. Estimates suggest that by 2050, Europe could meet 40 to 75% of its critical metal needs through recycling. At the same time, the development of « frugal » AI models, designed to be less resource-intensive, would reduce pressure on raw materials. These models would be lighter, more energy-efficient, and adaptable to simplified infrastructures. An Uncertain Future for Artificial Intelligence The question therefore remains whether artificial intelligence will be able to overcome this monumental obstacle. Political decisions in key countries such as China or the Democratic Republic of Congo could have profound ramifications for access to essential resources. Unexpected delays in metal deliveries could slow the development of crucial AI innovations. Vigilance and anticipation of supply-related risks are therefore essential to ensure the sustainability of these future technologies. Companies must prepare for a constantly evolving landscape marked by competition for these vital resources, while innovating to ensure their growth in the face of growing uncertainties.
