Microsoft Corp. greenhouse-gas emissions rose by 25 percent [1] in the most recent fiscal year due to the buildout of AI-driven data centers.
This surge highlights a growing tension between the rapid deployment of generative artificial intelligence and the corporate climate goals of the world's largest tech companies. As AI workloads demand more power, the physical infrastructure required to support them is creating a larger environmental footprint.
For the reported 2025 fiscal year, Microsoft emitted 20 million metric tons of CO2-equivalent [1]. A significant portion of that increase was tied to emissions from energy the company purchased or acquired to run its operations, known as Scope 2 emissions [3].
Company officials said that sustainability solutions are not scaling fast enough [2]. The rapid growth of AI workloads and the construction of new data centers across the company's global network outpaced its renewable-energy matching efforts [1, 2].
Despite the rise in total emissions, the company maintains it is pursuing green energy targets. "We matched our entire electricity consumption with renewable energy last year," a Microsoft spokesperson said [4].
This recent spike is part of a longer trend. Microsoft emissions have increased 23.4 percent since 2020 [5]. The company's global data-center network continues to expand to meet the processing demands of AI, which requires significantly more energy than traditional cloud computing tasks [1, 2].
“Microsoft’s greenhouse-gas emissions rose by 25 percent year-over-year.”
The disconnect between Microsoft's 100% renewable electricity match and its rising total emissions illustrates the gap between 'matching' energy and absolute carbon reduction. While the company buys enough renewable credits to cover its use, the actual physical demand of AI infrastructure is increasing the total volume of carbon released into the atmosphere, suggesting that current green energy procurement cannot fully offset the scale of the AI boom.
