The AI Race is Pushing Microsoft’s Sustainability Goals Further Away

What We’re Showing

A chart with Microsoft’s electricity consumption (in terawatt-hours) and their emissions (in million metric tons of CO2e) from FY 2020–23, as sourced from company reports. Microsoft’s financial year runs from July 1st to June 30th.

Key Takeaways

• Microsoft’s electricity consumption has doubled from 11 TWh to 24 TWh in four years.
• This is accompanied by a 42% increase in total emissions—indicating a relative growing share of renewable energy sources
• Both trends coincide with Microsoft Azure’s use to train and run AI models, of which OpenAI’s ChatGPT is the most prominent
• In fact, Microsoft spent “hundreds of millions of dollars” to develop a super computer just for ChatGPT, which involved linking thousands of Nvidia GPUs

More AI, More Energy, More Emissions

Training AI models requires a lot of compute. The data centers built to provide said compute are more power hungry than those providing traditional email or web services. In fact, the construction of these centers has accounted for 30% of the emissions increase between 2020 and 2023.