Microsoft reports a 25% rise in carbon emissions, attributing it to the growth of AI data centres

Microsoft’s push to expand its AI infrastructure is complicating the fulfillment of its climate commitments. According to a famous publication, the tech giant revealed that its carbon emissions increased by 25% in 2025, with the rapid expansion of AI data centres being the primary cause. These figures were published in Microsoft’s 2026 Sustainability Report, which indicates that the company’s emissions reached 34 million metric tons in 2025 “without specific interventions.”

According to Microsoft, the increase was “driven primarily by the expansion of our data centre infrastructure.” The company also noted that its decision last year to stop purchasing “unbundled, non-additional renewable energy certificates” contributed to the rise in emissions.

These figures pose a new challenge to Microsoft’s ambitious goal of becoming carbon-negative by 2030—a target by which the company aims to remove more carbon from the atmosphere than it emits. This is not the first setback; Microsoft’s previous sustainability report had already revealed an increase in emissions as the company accelerated its investments in AI infrastructure.

The AI boom makes meeting climate goals difficult

The report acknowledges the growing environmental cost of artificial intelligence. “While AI infrastructure is driving demand for energy, water, land, and materials, sustainability solutions are not scaling fast enough to meet that demand,” Microsoft stated. The results reflect a broader trend in the tech sector, where companies are investing billions of dollars in new data centres to support increasingly powerful AI models. These facilities require massive amounts of electricity to power thousands of specialised processors, as well as large-scale cooling systems to keep them operational.

Google also reported a 25% increase in supply chain emissions in its latest sustainability report, while Amazon recorded a 16% rise during the same period, highlighting the environmental impact of the AI race. Tech giants are investing billions in data centres.

The scale of investment in AI infrastructure has become even more apparent in recent months. Reuters recently reported that Meta plans to invest 13 billion Canadian dollars (approximately 9.17 billion US dollars) in its first data centre in Canada, located in Alberta. The facility will initially have a capacity of 1 gigawatt, with plans to expand to 1.8 gigawatts, and is expected to consume as much electricity as nearly 800,000 households.

To support the project, Meta states it will invest in new power generation, modernise the electrical grid, and offset the facility’s electricity consumption through investments in clean, renewable energy. The company also plans to use a closed-loop liquid cooling system to reduce water consumption.

Water usage has also become a growing concern as AI infrastructure expands. Amazon recently revealed that its global data centres consumed around 2.5 billion gallons of water in 2025, although the company noted that it reduced direct water usage compared to the previous year by utilising treated wastewater and outdoor-air cooling systems whenever possible.

As AI adoption accelerates, tech companies are increasingly trying to balance rapid infrastructure expansion with ambitious sustainability goals. Microsoft’s most recent report shows that, for now, keeping both objectives aligned remains a considerable challenge.