Big Tech's carbon footprint is surging as data center construction outpaces renewable energy development. Between 2025 and 2026, Google and Microsoft each reported a 25% increase in total carbon emissions, while Amazon saw a 16% rise. Collectively, these three giants emitted 19 million metric tonnes of carbon dioxide equivalent last year, a volume comparable to one-third of France's total annual output.

While industry defenders argue that AI will optimize energy efficiency across sectors, a study published in NPJ Climate Action disputes this optimism. Researchers found that if AI provides even marginal economic gains to the fossil fuel industry, the resulting "enabled emissions" will far outweigh any climate benefits. In a scenario where AI benefits both renewables and fossil fuels equally, global emissions could rise by up to 1.8 billion tons annually—an amount equivalent to the yearly carbon footprint of Mexico.

Achieving a net reduction in emissions would require the renewable energy sector to capture AI-driven efficiency gains at a rate four to five times higher than the fossil fuel industry. Given current global political and economic conditions, the authors suggest this outcome is unlikely. Instead, the integration of AI is positioned to reinforce fossil fuel incumbency, extending the industry's dominance and rendering current climate governance frameworks largely ineffective.