Google's emissions up 48% since 2019, driven by AI data centers

Google's emissions are up 50%

Google’s greenhouse gas emissions have grown dramatically, climbing 48% since 2019 according to its 2024 environmental report. The company now emits 14.3 million metric tons of carbon dioxide equivalent annually.

This surge directly contradicts Google’s 2020 pledge to achieve net-zero emissions across its operations by 2030. The primary driver is the immense energy consumption of its data centers, which power artificial intelligence and other cloud services.

AI is the primary culprit

The report explicitly links the increase to artificial intelligence. “This result is primarily due to increases in data center energy consumption and supply chain emissions,” Google states. The company notes the difficulty of forecasting the future environmental impact of AI.

Training and running large AI models requires vast computational power. As Google, Microsoft, and Amazon race to integrate AI into every product, their energy demands and associated emissions have skyrocketed.

Google’s data center electricity consumption grew 17% in 2023. The company argues it is working to mitigate this by pursuing carbon-free energy and improving efficiency, but these efforts are not yet offsetting the explosive growth in AI compute.

A broken climate promise

In 2020, Google CEO Sundar Pichai announced the company’s “third decade of climate action” with the goal of running on 24/7 carbon-free energy by 2030. The latest data shows it is moving in the opposite direction.

The company’s total emissions have risen year-over-year since 2020. Its 2024 report shows a 13% year-over-year increase from 2022 to 2023, a faster rate than the previous period.

This trend mirrors the broader tech industry. Microsoft’s emissions are up roughly 30% since 2020, largely due to AI infrastructure. Both companies are investing heavily in carbon capture and clean energy, but these are long-term bets against a short-term explosion in energy use.

The immense energy cost of AI

The computational hunger of AI systems is well-documented. Key factors driving data center energy use include:

• Model Training: Creating a single large language model can consume energy equivalent to the annual electricity use of hundreds of homes.
• Inference: The process of running a trained model to generate answers, which happens billions of times per day for services like Google Search and Gemini, requires continuous, massive compute.
• Hardware Manufacturing: Building the specialized servers and chips for AI has a significant carbon footprint in Google’s supply chain, which accounts for over 75% of its total emissions.

Google claims its AI-optimized data centers and Tensor Processing Units (TPUs) are more efficient than typical systems. However, this efficiency is being overwhelmed by the sheer scale of new AI deployment.

Can tech giants fix the problem?

Google maintains it is still committed to its 2030 goal. Its strategy rests on three pillars: improving energy efficiency, accelerating the deployment of carbon-free energy like solar and wind, and developing new carbon removal technologies.

“Our approach will continue to evolve and will require a balance between our climate commitments and the energy needs of our AI and digital infrastructure,” the report states. Critics argue this is an admission that the goals may be incompatible with current business plans.

The situation highlights a central tension for the tech industry. The promise of AI is driving unprecedented investment and consumption, but its environmental cost threatens to undermine corporate climate pledges and global emissions targets. Google’s rising emissions are a stark indicator that the AI boom has a major, and growing, carbon footprint.