The intense global competition to dominate artificial intelligence is entering a critical new phase. According to analysis by Vittorio Quaglione, the decisive advantage is no longer just about superior chips or algorithms. The new battleground is reliable and affordable electricity. As AI systems and the data centres that power them expand exponentially, the ability to supply continuous, massive amounts of power is becoming the ultimate chokepoint.
From Chips to Kilowatts: The Shifting Bottleneck
The rivalry has so far seen the United States leverage its dominance in chip design, strategically limiting exports to China. China, in turn, has used its control over rare-earth materials essential for manufacturing chips and advanced tech components as a counter-pressure point. However, this dynamic is changing. The sheer scale of AI's computing hunger means that all the data centres in the world are useless without a constant, affordable energy supply.
The International Energy Agency warns that roughly 20% of planned global data-centre capacity could be at risk by 2030 due to grid bottlenecks and connection delays. As energy supplies tighten, costs will inevitably rise, impacting both businesses and households. This shift in the competition's focus was anticipated by economist Albert O. Hirschman, who argued that an economy's true power lies in managing the choke points affecting its core industries.
Contenders on the Power Grid: China's Lead vs. US & EU Challenges
China appears to be pulling ahead in this energy-focused leg of the race. It has executed a massive build-out of energy supply and distribution infrastructure, with a strong emphasis on renewables like solar and hydropower. China installed a staggering 429 gigawatts of new power generation capacity in 2024 alone—more than six times the net capacity added in the United States. It can now add between 500 gigawatts and one terawatt of capacity annually.
Furthermore, China's industrial policy is tightly coordinated. To offset the cost of using domestic chips, local governments offer electricity subsidies of up to 50% for data centres that use them. This creates a powerful, integrated strategy linking chip consumption, data centre growth, and energy affordability.
In contrast, the United States faces significant structural challenges. While companies like OpenAI plan data centres requiring up to 10 gigawatts—comparable to New York City's summer peak load—the supporting energy infrastructure lags far behind. Building a data centre takes years, but completing the necessary transmission lines can take nearly a decade. This misalignment is causing strain; wholesale electricity prices near some major US data-centre hubs have skyrocketed by up to 267% in five years.
Europe, meanwhile, possesses unique strengths but struggles with execution. The EU develops over one-fifth of the world's clean energy technology and has a highly interconnected electricity grid. Its policy explicitly treats grids as a strategic asset. However, progress is slow. The average European grid project takes over ten years to complete, with half that time spent on permitting. Current investment projections are only 10-15% of what is needed, and over 500 GW of offshore projects are stuck in connection queues.
The Road Ahead: Energy as the Ultimate Decider
The conclusion is clear: the next phase of the AI race will be won or lost on the power grid. China's holistic strategy of boosting domestic chip production while aggressively expanding and subsidising its electricity infrastructure addresses every critical dimension. The United States, complacent in its lead in chip design and AI models, is failing to plan for the coming energy crunch. Europe has the technical know-how and policy vision to make AI's energy demand cleaner and more secure but may lack the institutional speed to capitalise on it.
For any nation or region—including Nigeria, which must consider its own energy stability for future tech growth—the lesson is unmistakable. To compete in the AI-driven future, you must first direct your energies toward energy.
