Generative AI could triple data center carbon emissions
New research finds that the rapid development of and demand for generative AI has accelerated greenhouse gas emissions from data centers.
A report from Morgan Stanley found that the data center industry is expected to emit 2.5 billion tons by 2030, three times more than forecasts had generative AI not been deployed.
The additional demand from GenAI is expected to increase emissions from 200 million tons this year to 600 million tons in 2030, mainly due to the construction of more data centers to meet the demand for cloud services.
Net zero targets
Morgan Stanley’s report outlined that 60% of the emissions figures will come from the operations of the data centers, as they require vast amounts of power to run. The remaining 40% will likely come from the carbon released during the production of the building materials and infrastructure for the centers.
Of Google already reports it A 48% increase in emissions over the past five years, unsurprisingly, calls net-zero emissions targets into question. The tech industry already accounts for 40% of total annual US emissions – so carbon removal technologies are poised to play a major role in meeting environmental goals.
The difficulty in reducing the environmental impact of data centers is that they can reduce energy consumption through water cooling systems, but it takes a huge amount of water to do so. As water becomes an increasingly precious resource, those systems must hinder the tech giant’s green goals and put enormous pressure on areas with ‘severe water shortages’.
There is uncertainty surrounding the future of AI and its impact on the environment. Carbon removal and carbon capture, utilisation and storage (CCUS) technologies are not yet fully developed. Morgan Stanley suggests that CCUS technology will require a $15 billion investment to bring it up to scratch. The research also points to reforestation projects as a potential tool for net zero goals in the future.
Via The register