The US-Israeli war on Iran is a wake-up call for the European Union: energy remains a critical strategic vulnerability. But addressing this vulnerability by weakening elements of the European Emissions Trading System, as the European Council seems to want, would do nothing to confront a challenge that extends well beyond dependence on imported fossil fuels. The real source of Europe’s vulnerability is an energy system that is fundamentally incompatible with economic power in the 21st century.

Geopolitics today is increasingly shaped by two inextricably linked forces: energy and AI. Data centres convert energy into compute, which forms the basis of economic and strategic power. AI, at scale, thus amounts to critical infrastructure, with the economies that transform energy into “intelligence” most efficiently gaining a decisive advantage.

The United States and China are currently fighting for the lead in this race, though their models diverge sharply. China’s model is based on a combination of system-level coordination – which recognizes the interconnections between energy, infrastructure and digital capacity – and intense micro-level competition.

Crucially, China is not only transforming its energy system; it is expanding it. Massive investment in solar, wind, storage and grids is driving down the marginal cost of electricity, creating favourable conditions for large-scale AI deployment. This expansion is bolstered by a financial architecture that channels low-cost credit into infrastructure and industry, accelerating cost reduction and broad adoption.

Chinese AI companies are already beginning to translate lower energy costs and increased system efficiency into dramatically cheaper compute. Firms like MiniMax and Moonshot reportedly charge around US$2 to US$3 per million output tokens – the basic units of AI-generated text and a standard measure of compute usage – compared with roughly US$15 for leading US models. This directly lowers the price of deploying AI at scale.

While this model remains far from perfect, as shown by grid bottlenecks, regional imbalances between generation and demand, and continued reliance on coal to stabilize supply, its underlying logic is strong. Beyond competing at the technological frontier, China is seeking to deploy AI capabilities as broadly as possible.

To this end, China is not exporting primarily raw compute, which others would have to transform into usable applications. Instead, it exports products in which digital intelligence is embedded, such as electric vehicles, industrial automation tools and telecommunications equipment. In other words, it exports electricity that has been transformed not just into intelligence, but into industrial capability.

The US, by contrast, is focused squarely on the technological frontier, with a small number of firms, supported by deep capital markets, investing at unprecedented scale in advanced chips, large models, and cloud infrastructure. While this model excels at breakthrough innovation, it also leads to high costs, concentrated capacity, and limited access to compute.

The US model is also highly ( and increasingly ) energy-intensive. The Lawrence Berkeley National Laboratory estimates that data centre demand could grow from 176 terawatt hours ( TWh ) in 2023 to 325-580 TWh by 2028, amounting to 6.7-12% of total US electricity consumption. Already, power availability is becoming a binding constraint.

The US and China models have one thing in common: both reflect the recognition that AI is now a key determinant of economic power. What China sees more clearly is that abundant energy is essential to AI adoption.

Europe seems to be missing both lessons.

Given Europe’s longstanding ambition to lead the global green transition, one might imagine that its position should be promising, at least on the energy side of the equation. But this ambition has been consistently undermined by fragmentation, regulatory uncertainty, and the absence of a coherent industrial strategy that emphasizes the provision of abundant, low-cost electricity. As a result, expansion of the energy supply, particularly in renewables, grids and storage, has been insufficient, and electricity prices remain persistently higher than in competing economies.

But this is not just a cost problem; it is a structural threat to European competitiveness. As Europe misses its climate targets, it will also continue to fall behind in the integration of energy and digital capacity. This will leave it dependent on the US for frontier AI, and on China for cost-efficient industrial deployment. The Ukraine war, US tariffs, and Chinese export controls have already shown Europe how risky such dependencies can be.

While Europe cannot replicate the Chinese model, it can internalize its central lesson: the energy transition is not only about sustainability, but also about scale, cost, and industrial transformation. To do so, Europe will have to overcome formidable institutional constraints. Interestingly, the challenge of coordinating a large, heterogeneous system – aligning infrastructure, financing, and local incentives – is shared by both China and the Europeans.

But Europe faces additional problems: fragmented capital markets, state-aid rules and limited fiscal capacity can slow down investment and prevent projects from reaching the scale necessary to deliver tangible cost improvements. The overall volume of capital is only part of the challenge. To have an impact, investment must be sufficiently coordinated and targeted.

This requires, for starters, de-risking mechanisms, long-term contracts and predictable regulatory frameworks that support more effective allocation of private capital. Moreover, EU-level instruments should assume a central role in crowding in investment, thereby helping to overcome fragmentation. Finally, in cases where projects have clear strategic value, there should be greater flexibility on fiscal and state-aid rules. The objective is not to choose between state and market dominance, but rather to design a framework in which public coordination and private capital reinforce each other.

Europe will never have the abundant fossil fuel resources of the US, but by ramping up investment in its energy infrastructure, and organizing that capital effectively, the EU can achieve the energy diversification and cost reductions seen in China. This would protect Europe from energy price spikes like those triggered by the Iran war. It is also a prerequisite for the EU to be able to compete in the defining economic race of our time: building the ability to turn electricity into intelligence.

Lucrezia Reichlin is a professor of economics at the London Business School and a former director of research at the European Central Bank.

Copyright: Project Syndicate