Data-center emissions to triple by 2030, thanks to generative AI
The rapid advancement and increasing demand for generative artificial intelligence (AI), are significantly contributing to greenhouse gas (GHG) emissions from data centers, according to recent research. A study by Morgan Stanley predicts that the data center industry is on track to emit 2.5 billion tons of GHGs by 2030. This figure is three times higher than what was previously projected without the influence of generative AI.
Generative AI's demand surge to boost emissions
The surge in demand for generative AI is expected to increase emissions from 200 million tons this year, to a staggering 600 million tons by 2030. This rise is primarily attributed to the construction of additional data centers needed to meet growing cloud service demands. Morgan Stanley's report reveals that 60% of these emissions will be generated by the data centers' operations due to their high power consumption.
Data center construction and net-zero targets
The remaining 40% of emissions are expected to come from carbon released during the production of construction materials, and infrastructure for these data centers. This revelation comes as Google has reported a 48% increase in its emissions over the past five years, casting doubt on the feasibility of achieving net-zero emission targets. The tech industry currently accounts for a significant 40% of all annual US emissions, highlighting the potential role of carbon dioxide (CO2) removal technologies in meeting environmental goals.
Environmental impact and water scarcity
Mitigating the environmental impact of data centers presents a unique challenge. While energy consumption can be reduced through water-cooling systems, these require vast amounts of water, making them less viable in regions with high water scarcity. This issue further complicates tech giants' efforts to achieve their green goals. The future impact of AI on the environment remains uncertain due to these complexities and challenges.
Carbon capture technologies need significant investment
Morgan Stanley suggests that carbon capture, utilization, and sequestration (CCUS) technologies require a substantial $15 billion investment to meet necessary standards. The study also identifies reforestation projects as potential tools for achieving net-zero targets in the future. These findings underscore the urgent need for innovative solutions and significant investments to address the environmental impact of generative AI's rapid development and increasing demand.
