Google promised AI would help cut 5–10% of global emissions by 2030 — its own 2024 report disclosed AI drove the company's emissions up 48% since 2019
"AI has the potential to mitigate 5–10% of global greenhouse gas emissions by 2030 — the equivalent of the annual emissions of the European Union — through optimization of energy grids, transportation routing, and climate-modeling acceleration."
Greenhouse gas emissions in 2023 totalled 14.3 million tonnes CO2e, an increase of 48 percent compared to 2019, primarily driven by increased data center energy consumption and supply chain emissions associated with AI infrastructure build-out. — Google 2024 Environmental Report (published July 2024), July 2024. source ↗
That's roughly 893,750 US suburban households' annual carbon.
Pichai's COP28-timed framing positioned Google's AI as a climate-solution multiplier targeting 5–10% of global emissions reductions. Eight months later, Google's own published Environmental Report disclosed that their AI build-out had moved their corporate emissions in the opposite direction: 14.3 million tonnes CO2e in 2023, up 48 percent against the 2019 baseline. At the platform's declared baseline of 16 tonnes CO2e per US suburban household per year, Google's annual corporate emissions equal the household carbon footprint of ~894,000 US suburban households. The report itself attributes the increase to data center energy demand and supply-chain emissions from AI infrastructure — the same infrastructure cited as the mechanism for the promised global cut.
Baseline: 2024 US suburban household consumption: 10,500 kWh electricity / 109,500 gal water / 16 tonnes CO2e per year. · Methodology · How this is verified