Google has cut its own operational greenhouse gas emissions by 2% in 2025 despite a dramatic rise in electricity use as it continues to rapidly scale up its artificial intelligence (AI) infrastructure, according to its Environmental Report. Overall, electricity demand increased 37% for the company in the year, driven by unprecedented demand for electricity related to AI computing, and the acquisition of clean energy projects reduced the environmental footprint.

The report points out the difficulty that tech firms are having with scaling up AI and climate commitments.The report notes the increase in their struggles to scale up AI while maintaining climate obligations. In total, Google inked contracts for over 12 gigawatts (GW) of new clean energy generation in 2025, the highest volume of renewable energy procurement for the year. When working, the projects would generate enough electricity to power a country the size of Greece for an entire year, bringing Google to the top tier of large-scale corporate clean energy purchasers in the world.

Renewable energy investment helps mitigate electricity demand growth.Investment in clean energy contributes to mitigating electricity demand growth.

The new deals are part of a larger effort to reduce the environmental footprint of Google's activities and to meet the growing demand for energy in the AI sector, Google said. From 2010 to 2025, the company signed over 240 clean energy purchase agreements corresponding to almost 35 GW of new renewable energy capacity.

The report says that the projects will ultimately provide power for over 28 million homes in the United States. The company also continued to buy 100% of the electricity it uses each year from renewables for the ninth consecutive year.

Google said electricity consumption by its data centres had risen to the highest it has ever seen, but the centres are still running efficiently. The company has 83% lower overhead energy consumption than the industry average for its data centre infrastructure, which is expected to help to lower emissions despite the continued growth of computing workloads.

AI introduces new challenges for the climate.AI poses new climate challenges.

The company said it recognizes the swift proliferation of AI infrastructure has introduced new environmental issues. The company said its growth is now faster than the rate of decarbonisation of electricity grid, making progress toward long-term climate targets more difficult.

Google has reaffirmed its dedication to reducing its environmental impact and advancing AI technologies in the report. To meet its climate targets, it will need to continue investing in ample and cost-effective clean energy, as well as technologies that can reduce emissions in its own operations and industry more broadly," it added.

Growing beyond renewable energy.

In order to meet future electricity demands, Google announced that it is investing in the generation of clean energy on the local electricity grids for its data centres. The company is also advancing technologies like nuclear power, enhanced geothermal power, and long-term development of nuclear fusion.

The report suggests that corporate energy strategies are shifting from the conventional renewable energy certificates. Long-term electricity purchase agreements, utility partnerships, and investments in new generation capacity are among the energy market strategies that are becoming commonplace among large technology firms to improve electricity grids and help them reach sustainability goals.

Google has said it designs its power deals to account for the actual price of the electricity it uses and wants to "facilitate the availability of reliable, affordable and cleaner energy for the communities in which it operates.

AI Tools have Environmental benefits.

In addition to its own emission reduction efforts, Google emphasized the importance of AI-driven technologies in its worldwide climate and environmental efforts.

The firm calculated the reductions in CDEs avoided by 2025 were well over 58 million metric tonnes of carbon dioxide equivalent, attributable largely to the efficiency of equipment, optimisation of software and increased computing power, as well as clean energy procurement. Otherwise, Google said its total carbon emissions footprint linked with its growth ambitions would have been around five times as much.

It was also estimated that, during the year, users could have avoided approximately 41 million metric tonnes of CO2 equivalent emissions through the use of nine of Google's products and services, among them Google Earth, Google Fusion, Google Earth for Education, and Google Sniper. These included tools to help with solar and wind energy planning in Google Earth, energy-saving capabilities of Nest smart thermostats, and fuel-efficient routing in Google maps.

Google also highlighted AI applications for disaster preparedness—such as systems that enhance wildfire, flood, earthquake and extreme weather detection. The company announced in the field of biodiversity conservation the development of Perch, a bioacoustic AI model that analyzes large volumes of nature data, and SpeciesNet, an image recognition model that recognizes more than 2,000 animal species from photos of wildlife captured by cameras with over 94% accuracy.

Carbon emissions in Supply Chain keep climbing.

While Google has made strides to lower emissions from its direct operations, overall, its carbon footprint expanded in one significant area. Emissions from the supply chain increased by 25% in 2025, primarily because of increased demand for AI infrastructure and manufacturing activity in the Asia-Pacific region, where reliance on fossil fuels in the electricity grid remains high.

The company also said that shortage of land, increasing construction costs and policy-related challenges were impacting the deployment of clean energy in its supply chain.

The uptick reflects a persistent difficulty in controlling Scope 3 emissions, all of which are considered the biggest challenge for large tech companies. As the AI infrastructure continues to grow across the globe, these indirect emissions are drawing the interest of investors and regulators alike.

While Google claims its new findings prove that emissions from operations can go down even in the face of fast-growing AI, emissions from its wider supply chain are still a huge problem. As AI adoption expands, the company stated that ongoing cooperation between energy systems, the development of infrastructure and innovation in the field of technology will be crucial to help meet long-term climate objectives.