The global conversation around artificial intelligence is largely focused on models, data centres and technological breakthroughs.

Yet behind every AI query, every large language model and every new digital service lies a less visible reality: energy.

According to Robert Friedland, founder of Ivanhoe Mines and a long-time investor in critical resources and technology, Europe risks underestimating the scale of the challenge. While governments and companies accelerate investments in AI, the continent must also confront a crucial question:

Who will generate the electricity needed to power the AI economy ?

AI’s invisible dependency

Artificial intelligence is often described as a digital revolution. In reality, it is also a physical one.

Data centres require electricity. Semiconductors require critical minerals.

Cloud infrastructure requires vast amounts of power and cooling capacity.

“Artificial intelligence is a ludicrously large consumer of electrical energy,” Friedland warned during his opening address at Nexus Luxembourg.

The observation may sound obvious, but it highlights a growing strategic concern.

As AI adoption accelerates, demand for electricity is expected to surge. Every new model, every AI-powered search and every autonomous system increases pressure on energy infrastructure.

The challenge is no longer simply about innovation.

It is about whether Europe can generate enough power to sustain its ambitions.

A new form of strategic dependency

For decades, Europe’s energy security was discussed primarily through the lens of oil and natural gas. Today, a new dependency is emerging: computing power.

The continent’s AI ambitions increasingly rely on data centres, cloud infrastructure and advanced computing facilities. All of these require stable, affordable and abundant electricity.

Without it, Europe risks becoming dependent on regions that control the infrastructure powering artificial intelligence.

For Friedland, the issue extends beyond economics.

It is a matter of sovereignty. The same geopolitical tensions that have exposed vulnerabilities in energy markets are beginning to affect technology supply chains.

Conflicts in Ukraine and the Middle East have demonstrated how quickly global dependencies can become strategic weaknesses.

“The world economy is being Balkanised by war,” he argued. In such an environment, technological leadership cannot be separated from energy security.

The risk of falling behind

The United States and China are investing heavily in artificial intelligence infrastructure.

Massive capital is flowing into data centres, semiconductor manufacturing and energy production. Europe, meanwhile, faces a more complicated equation.

The continent wants to lead in sustainable technology and AI innovation while simultaneously reducing carbon emissions and strengthening energy resilience.

Those objectives are compatible but they require significant investment.

According to Friedland, Europe cannot assume that access to electricity will simply follow demand.

It must actively build the infrastructure needed to support the next generation of digital industries.

“Europe needs a better way to generate electrical energy if it wants any hope of developing AI,” he said. The warning is particularly relevant as hyperscale data centres continue to expand across the continent.

Geothermal energy and the search for alternatives

One solution highlighted by Friedland is geothermal energy. Unlike solar and wind power, geothermal resources provide continuous electricity generation regardless of weather conditions.

For AI infrastructure, which requires reliable baseload power around the clock, that characteristic is especially attractive.

Friedland argued that the Earth itself represents a vast energy reservoir that remains largely underutilised.

By drilling deeper and deploying new technologies, countries could access a stable and low-carbon source of electricity capable of supporting data centres and industrial activity.

Whether geothermal becomes a major pillar of Europe’s energy mix remains uncertain.

However, the broader message is clear.

The AI transition will require an energy transition as well.

Lessons for Luxembourg

The argument resonates in Luxembourg. The country has built a reputation for identifying strategic opportunities early, most notably through its investments in satellite infrastructure decades ago.

At the time, the decision appeared ambitious. Today, it is widely regarded as visionary. Friedland suggested that Europe faces a similar moment.

Artificial intelligence will not only reshape software and services. It will reshape infrastructure, energy systems and industrial policy.

Countries that understand this connection early may secure a lasting competitive advantage.

Those that do not risk becoming consumers rather than producers of the next technological revolution.

Beyond algorithms

The AI debate is often dominated by discussions around models, productivity and automation.

Those questions matter but Friedland’s intervention served as a reminder that the future of artificial intelligence may be decided by factors far removed from Silicon Valley.

Electricity generation.Critical raw materials. Energy resilience. Infrastructure investment.

In the coming years, Europe’s ability to compete in AI may depend less on developing the smartest algorithms than on securing the resources needed to run them.

The race for artificial intelligence is increasingly becoming a race for energy independence and that race may prove just as important.