Welcome to the Digital Download, where I combine my studies of Computer Science and Philosophy to provide insight into the rapidly-advancing world of AI, and the impact it is having on the energy sector. With issues ranging from privacy and job security to geopolitics and overlooked environmental impacts, debate around the correct usage of AI tools is gaining volume every day, and this column aims to bring you all the most recent and important developments in the discussion.
Data Centre Diplomacy: AI on the Global Stage
ChatGPT: a lecturer’s sworn enemy and a stressed student’s saving grace. While many of us are familiar with the complexities of generative AI, how many are aware of the geopolitical turmoil sparked by its rapid growth?
AI has remained in the spotlight since its popularisation a few years ago, but recent literature has begun to highlight a critical issue with its development: power. Asking a Large Language Model (LLM) like ChatGPT a single question consumes 100 times more energy than the average Google search, and training such a model can emit up to 284,000kg of CO2. This equals the emissions generated by five cars over their entire lifespan. As the demand for AI infrastructure grows, energy consumption is projected to increase, with data centres accounting for the largest share of this rise. As a result, the AI arms race is shifting its focus to these data centres, as nations compete for a stake in what is increasingly being referred to as the second industrial revolution.
Currently, the United States hosts the largest share of global data centres (49%), placing it in a powerful position but also straining its already overburdened energy grid. A key characteristic of the AI industry is its reliance on a global network of commercial partners, including Chinese technologies, Taiwanese semiconductors and American software. As the demand for data centres continues to grow, the geographical location of these facilities is set to become a critical factor in technological progress and competition. Amid the most acute geopolitical turbulence since the Cold War, global alliances are shifting, with both China and the West seeking to reduce their dependence on foreign tech resources. In this context, the countries that succeed in hosting the data centres of major tech giants will gain significant economic, political and technological advantages.
This power struggle is exacerbated by the increasingly ambitious sustainability commitments of companies such as Amazon and Google. If the United States aims to continue decoupling its economy and make it more difficult to purchase from China, it must develop a robust and commercially attractive alternative. This would require scaling up its energy capacity, accelerating the transition to renewable energy, or forging partnerships with strong competitors. Canada and Nordic states offer several advantages, including leadership in green energy, high levels of connectivity, cooler climates that mitigate data centre overheating, and strategic alliances (e.g through NATO). For tech companies less focused on their carbon footprint, China and the Sunni Gulf states remain attractive options, offering cheap coal and oil as well as regulatory flexibility.
This next phase of the AI revolution will be critical, with geography and power at its core. This stage holds the potential to accelerate the energy transition, though it may also lead to a short-term increase in fossil fuel consumption to meet the surge in demand. The political stakes are equally high, as the RAND Corporation notes, “Nations across the globe could see their power rise or fall depending on how they harness and manage the development of AI”. Perhaps ChatGPT itself could offer a solution?