The rapid expansion of data centers around the world shows no signs of slowing down, as major technology companies explore innovative ways to support the burgeoning artificial intelligence (AI) sector. This exploration includes options such as nuclear power, liquid cooling systems, and advances in quantum computing.
Despite the enthusiasm, critics warn that the slowing pace of energy efficiency improvements highlights the hidden environmental costs associated with technology sector growth. Somya Joshi, Head of Division: Global Agendas, Climate and Systems at the Stockholm Environment Institute, pointed out that the real environmental impact is often obscured by the urgent need for tech companies to deliver products quickly.
As reported by the International Energy Agency, investments in data centers are expected to increase, driven by increasing digitalization and the rise of generative artificial intelligence. This projected demand raises concerns about a substantial increase in electricity consumption and the environmental implications of AI technologies.
Raj Hazra, CEO of Quantinuum, noted that every major technological advance goes through a period of reflection and reevaluation, suggesting that the current enthusiasm for generative AI may lead to challenges in the future.
Data centers are critical to the infrastructure that supports modern cloud computing and artificial intelligence applications, and their energy consumption continues to increase. Giampiero Frisio, ABB's president of electrification, highlighted significant growth in the company's data center business, predicting growth of more than 24% in 2024. He advocates improving energy efficiency through existing technologies such as the HiPerGuard medium voltage continuity system, which can support large systems with continuous power supply.
Frisio also pointed to liquid cooling as a promising solution for improving energy efficiency, especially as server power density increases. He predicts that modular nuclear systems will become a viable option in the next five to 10 years.
Big Tech's foray into nuclear energy has gained momentum, with companies including Microsoft, Google and Amazon securing multibillion-dollar deals to improve their energy capabilities for running complex artificial intelligence models. This aligns with a broader push toward efficient cooling solutions, such as liquid cooling, which uses water to maintain optimal temperatures for servers.
Schneider Electric recently acquired a controlling stake in Motivair Corp, a company specializing in liquid cooling technologies, as part of its strategy to strengthen its data center offerings. The CEO described the acquisition as seamlessly integrated into the company's plans, signaling a commitment to innovative cooling solutions.
In addition to nuclear and liquid-cooling technologies, some industry leaders argue that advances in artificial intelligence could help decarbonize data centers. Former Google CEO Eric Schmidt has suggested that investing in artificial intelligence could be key to addressing environmental challenges. However, Joshi firmly rejected this notion, calling it a familiar narrative that oversimplifies the complexities of sustainability within planetary limits.
Hazra also expressed concern about the sustainability of generative AI and the infrastructure it requires, highlighting the need for responsible resource management. He highlighted that quantum computing could play a vital role in creating sustainable solutions to AI challenges, highlighting its potential in handling the demands of modern computing frameworks.
As the computing landscape evolves, Hazra believes that companies will increasingly look to a combination of high-performance computing, artificial intelligence and quantum technology to meet their operational needs. The emphasis will shift not only to the availability of solutions, but also to ensuring that those solutions are sustainable and responsible, reflecting a growing awareness of environmental impact and resource efficiency in the technology sector.