Growth in the global data center industry will see emissions equivalent to 2.5 billion metric tons of carbon dioxide generated by the end of 2030, according to Morgan Stanley research.

GettyImages-167231386 carbon accounting smoke emissions.jpg
– Getty Images

The report said the hyperscalers, such as Google, Microsoft, Meta, and AWS, are all culprits in driving the proliferation of data centers as they expand their AI and cloud computing capabilities.

This is despite these companies pledging to slash global warming emissions from their data centers by the end of the decade.

The report added that, by 2030, global data center greenhouse gas emissions will amount to around 40 percent of what the US emits in a year.

However, it argued that this is set to create a large market for decarbonization solutions, with the build-out of data centers simultaneously increasing investments in clean power development, energy-efficient equipment, and so-called green building materials.

Many hyperscalers have made attempts to reduce carbon emissions, including via the purchase of clean energy and the deployment of new technologies.

Microsoft is one of the largest corporate buyers of renewable energy in the world – the company previously told DCD it has brought 34GW of capacity to the grid.

Furthermore, 2024 has seen the company sign PPAs in Ohio, Ireland, India, and France, as well as two PPAs in Texas totaling more than 800MW with Leeward Renewable Energy and RWE. It has also signed a 500MW distributed solar deal across the US with Pivot, and wind and solar deals in Europe with Repsol and European Energy.

Microsoft, alongside Google, Meta, and AWS has also recently agreed to participate in projects with the Open Compute Foundation to trial the use of “green concrete” for data center construction.

This month, Meta said it had launched a metric to provide real-time views of servers’ embodied emissions. The new metric adds a kind of depreciation schedule for server hardware’s carbon emissions measured against its utilization.