Google’s carbon emissions have surged by 51% since 2019, as the energy demands of artificial intelligence (AI) outpace the company’s efforts to decarbonize. The tech giant reported a 27% year-on-year increase in electricity consumption in 2024, driven by the expansion of datacentres required to power AI models like Gemini and GPT-4. Despite investments in renewable energy and carbon removal, Google’s scope 3 emissions—those tied to its supply chain—rose by 22% last year, underscoring the challenges of curbing emissions in an AI-driven economy.

The International Energy Agency (IEA) projects that global datacentre electricity demand could reach 1,000 terawatt-hours (TWh) by 2026, equivalent to Japan’s total consumption. AI alone is expected to account for 4.5% of global energy generation by 2030, according to research firm SemiAnalysis. This "non-linear growth" in energy demand complicates emissions forecasting, as AI’s rapid evolution makes future needs harder to predict.

Google has signed over 170 clean energy agreements since 2010, totaling 22 gigawatts (GW), with 8GW added in 2024 alone. However, the company acknowledges that emerging technologies like small modular reactors (SMRs) and advanced geothermal are lagging behind schedule due to high costs and regulatory barriers. "Getting to net-zero by 2030 will be very difficult," the company admitted in its report.

Despite these challenges, Google sees AI as a potential climate solution. The company aims to use AI to help partners reduce 1 gigaton of carbon emissions annually by 2030, through applications like energy-use prediction and solar potential mapping. "The emissions reductions enabled by AI could outweigh its own footprint," the report noted. Meanwhile, Google has already met its 2025 goal of eliminating plastic packaging for new products.

The findings highlight the tension between AI’s promise and its environmental costs, as tech firms grapple with balancing innovation and sustainability.