A new report from Morgan Stanley suggests that the rapid development of generative AI (GenAI) has accelerated the rate of greenhouse gas emissions from data centers. The report predicts that by 2030, data centers will emit 2.5 billion tons of carbon dioxide, three times higher than previously predicted without GenAI’s impact.
The increased demand for cloud services driven by GenAI is expected to lead to a rise in emissions from 200 million tons this year to 600 million tons by 2030. The majority of these emissions come from the operations of data centers, which require massive amounts of power to run. The remaining 40% comes from the carbon emitted during the manufacturing of construction materials and infrastructure for the centers.
This increased emission rate calls into question net-zero targets, especially considering Google’s reported 48% increase in emissions over the last five years. The tech industry already accounts for 40% of annual US emissions, making it crucial to develop technologies that can remove carbon dioxide from the atmosphere.
One challenge in mitigating the environmental impact of data centers is their reliance on water-cooling systems, which require enormous amounts of water. This strain on areas with high water scarcity hampers tech companies’ green goals and highlights the need for more efficient solutions.
While there is uncertainty around the future of AI’s environmental impact, researchers are exploring carbon removal technologies that can help achieve net-zero targets. Morgan Stanley suggests a $15 billion investment is needed to bring these technologies up to standard. Reforestation projects may also play a role in achieving these goals.
Source: https://www.techradar.com/pro/generative-ai-triples-the-carbon-dioxide-emissions-from-data-centers