Google has managed to decrease emissions from its data centers by 12% in 2024, a significant achievement considering a 27% surge in electricity consumption driven by the escalating demands of artificial intelligence. This reduction, detailed in the company’s 10th annual environmental report, underscores a concerted effort to decouple AI-related energy growth from a corresponding rise in carbon footprint, a pressing challenge for the technology sector.
The core of this accomplishment lies in a multi-pronged strategy that combines aggressive procurement of clean energy, innovations in hardware efficiency, and advanced infrastructure management. By signing a record volume of clean energy contracts and pushing the boundaries of computing efficiency, Google is actively working to mitigate the environmental impact of its operations while simultaneously leveraging AI as a tool to promote broader energy efficiency and sustainability.
Record Expansion of Clean Energy Portfolio
In 2024, Google made its largest-ever annual commitment to clean energy, signing contracts for over 8 gigawatts of new generation capacity. This figure is double the volume contracted in the previous year and brings the company’s total contracted clean energy since 2010 to more than 22 gigawatts. These agreements are projected to generate nearly four times more electricity than the company’s incremental load growth from 2023 to 2024, once operational. The company’s clean energy purchasing is estimated to have avoided over 8.2 million metric tons of carbon dioxide equivalent emissions in 2024 alone.
For the eighth consecutive year, Google maintained a 100% renewable energy match on a global basis. More importantly, it increased its hourly carbon-free energy percentage from 64% to 66%. This metric tracks whether its data centers and offices are matching their electricity consumption with clean power generated on the same grid in the same hour, a more rigorous standard for decarbonization. Nine of its 20 grid regions with data centers now achieve at least 80% carbon-free energy on this hourly basis.
Pioneering Advanced Energy Technologies
Next-Generation Nuclear and Geothermal
Looking beyond traditional solar and wind, Google is venturing into novel energy sources to ensure a consistent, 24/7 supply of clean power. The company signed the world’s first corporate agreement to purchase nuclear energy from small modular reactors (SMRs), partnering with Kairos Power. This deal will add up to 500 megawatts of clean energy to U.S. grids by 2035. SMRs represent a significant departure from conventional nuclear plants, offering a smaller, factory-built design that can be assembled on-site. The Kairos design utilizes a molten salt cooling system and ceramic pebble fuel, enhancing safety by allowing the reactor to operate at low pressure.
Google also significantly expanded its partnership with Fervo Energy for enhanced geothermal projects in Nevada. A new 115-megawatt project will increase the company’s enabled geothermal generation by nearly 25 times compared to its initial pilot project that came online in 2023. Enhanced geothermal systems create underground reservoirs to produce steam and generate electricity, providing a reliable source of carbon-free power.
Innovations in Data Center Efficiency
Hardware and infrastructure efficiency are critical components of Google’s strategy. The company’s data centers now deliver over six times more computing power for the same amount of electricity compared to five years ago. A key metric, power usage effectiveness (PUE), which measures the ratio of total facility energy to IT equipment energy, dropped to 1.09 in 2024, its best performance in six years. A PUE of 1.0 would represent perfect efficiency.
Custom-Built AI Hardware
Central to these efficiency gains are Google’s custom-designed Tensor Processing Units (TPUs), which are optimized for AI and machine learning workloads. The seventh-generation TPU, known as Ironwood, is nearly 30 times more efficient than the first Cloud TPU released in 2018. The sixth-generation Trillium TPU offers a 67% improvement in energy efficiency over its predecessor and delivers 14 times more compute per watt than the first-generation hardware. These specialized processors allow Google to handle the massive computational demands of AI with significantly less energy.
AI as a Tool for Global Emissions Reduction
Beyond reducing its own operational footprint, Google is applying its AI technology to help others decrease their environmental impact. According to its report, five of its AI-powered products enabled customers and users to collectively reduce an estimated 26 million metric tons of greenhouse gas emissions in 2024. This figure surpasses Google’s own total emissions of 11.5 million metric tons for the year. For example, Nest thermostats helped save over 25 billion kilowatt-hours of energy, avoiding approximately 8.4 million metric tons of emissions. Additionally, fuel-efficient routing in Google Maps is credited with preventing over 2.7 million metric tons of emissions, which is comparable to removing about 630,000 gasoline-powered cars from the road for a year. The company emphasizes that AI’s potential to help manage emissions in crucial sectors like transportation and energy is transformational.