The rapid construction of artificial intelligence data centers is pushing the carbon emissions of Microsoft, Amazon and Google sharply higher, complicating the companies’ promises to reach net-zero or carbon-negative operations.
The three technology giants collectively reported approximately 119 million metric tons of carbon dioxide-equivalent emissions in their latest annual disclosures, up from about 101 million metric tons one year earlier. That represents an increase of nearly 18%, with the companies linking much of the rise to data center construction, electricity use, equipment manufacturing and expanding global supply chains.
For comparison, France recorded approximately 369 million metric tons of domestic greenhouse gas emissions in 2024, excluding land use and forestry. The three companies’ combined total is therefore equal to roughly one-third of France’s annual territorial emissions. However, the comparison is not exact: corporate totals include emissions across international suppliers and business activities, while France’s figure measures emissions produced within the country.
Microsoft reported the largest annual percentage increase among the three companies. Its emissions climbed by approximately 25% to around 20 million metric tons, with the company saying the expansion of data center infrastructure was a primary factor.
TRENDING TODAY
Microsoft acknowledged that the growth of AI is increasing demand for electricity, water, land and construction materials faster than some environmental solutions can currently be deployed. The company continues to pursue its goal of becoming carbon negative by 2030, meaning it intends eventually to remove more carbon from the atmosphere than it emits.
Google’s emissions increased by approximately 18%, according to its latest environmental reporting. The company attributed much of the rise to supply-chain activity supporting the expansion of its business, including the construction of data centers and production of servers and other technical equipment.
Google also says its products and clean-energy investments can help reduce emissions elsewhere. Its estimates of emissions avoided through products such as fuel-efficient routing, smart thermostats and energy tools are based on internal methodologies and include uncertainty, according to the company’s own disclosures.
Amazon reported an approximately 16% increase in its carbon footprint. The company’s latest sustainability report describes AI as creating significant new demand for infrastructure, energy and water, even as Amazon continues investing in renewable power, more efficient computing equipment and lower-carbon construction materials.
Amazon remains committed to reaching net-zero carbon emissions by 2040 under its Climate Pledge. The company argues that more efficient custom chips and data center designs can reduce the amount of energy needed for individual workloads, although continued growth in total computing demand may still push overall emissions higher.
Much of the environmental impact occurs before a data center begins operating. Manufacturing steel, cement, computer chips, cooling systems and backup generators creates substantial supply-chain emissions. Once operational, the facilities require continuous electricity and may need significant amounts of water for cooling.
This helps explain why buying renewable electricity does not automatically eliminate a company’s entire carbon footprint. Clean-energy agreements can reduce operational emissions, but they do not necessarily remove the carbon generated by construction, equipment production or suppliers operating in countries where electricity remains heavily dependent on fossil fuels.
The challenge is likely to become larger as companies compete to build more computing capacity. JLL projects that global data center capacity could nearly double to approximately 200 gigawatts by 2030, driven mainly by hyperscale cloud services and AI.
The International Energy Agency expects global data center electricity consumption to roughly double by 2030, reaching about 945 terawatt-hours and accounting for just under 3% of worldwide electricity demand. Electricity use by data centers rose approximately 17% in 2025 alone, according to the agency.
The construction boom may also affect households and businesses far beyond the technology sector. Large data centers can require new power plants, transmission lines and water infrastructure. Depending on local regulations, some of those costs may be reflected in electricity bills, public subsidies or tax incentives.
At the same time, data centers can create construction jobs, local tax revenue and demand for new energy projects. The central policy question is whether technology companies, utilities or ordinary ratepayers should finance the infrastructure needed to support rapid AI expansion.
The companies argue that AI itself can contribute to climate solutions by improving power-grid management, weather forecasting, transport efficiency and industrial processes. Critics respond that those potential benefits should not be used to minimize the measurable emissions being created by today’s infrastructure buildout.
Why It Matters
The latest figures show that Big Tech’s AI ambitions are moving faster than its ability to reduce emissions. That gap could make corporate climate promises harder to meet and increase pressure on electric grids, water supplies and communities hosting large data center projects.
For consumers and businesses using cloud services, the reports also demonstrate that digital products are not environmentally weightless. Their footprint may be concentrated inside remote facilities and international supply chains rather than on the devices used to access them.
What Comes Next
Microsoft, Amazon and Google are expected to continue investing heavily in AI infrastructure while expanding renewable energy, nuclear power, advanced cooling and lower-carbon construction projects.
Regulators and local governments may increasingly demand clearer information about electricity use, water consumption, emissions and who will pay for new infrastructure before approving additional data center developments.
Communities across the United States are increasingly challenging large data center projects over their local impact.
More than 130 billion dollars in data center projects have been blocked or delayed this year.
Across the United States residents and local officials are increasingly resisting the construction of massive facilities required to support the rapid growth of artificial intelligence.… pic.twitter.com/xDieZWWCrt
— Massimo (@Rainmaker1973) July 11, 2026
