谷歌的最新可持续发展报告显示，由于人工智能（AI）的快速发展，其碳排放量再次上升。2023年，谷歌的“雄心排放”增加了11%，达到1150万吨二氧化碳当量，与2019年相比增加了51%。尽管谷歌的数据中心碳排放有所下降，但整体排放量仍未达到2030年减半的目标。报告指出，AI发展、气候政策变化、碳中和能源技术部署缓慢等外部因素，增加了实现气候目标的难度。文章强调了AI对能源需求的推动，以及应对气候变化所面临的挑战。

🏭谷歌2023年的“雄心排放”增加了11%，达到1150万吨二氧化碳当量，与2019年相比增加了51%。这使得谷歌距离其2030年将温室气体排放量减半的目标越来越远。

💡谷歌的数据中心碳排放减少了12%，但由于AI的发展，能源使用量持续上升。尽管有所下降，但与2019年相比，数据中心的碳排放仍然较高。

🌍谷歌认为，实现气候目标面临多重挑战，包括AI的快速发展导致未来能源需求难以预测、气候和能源政策的变化、碳中和能源技术的部署速度较慢以及某些市场缺乏碳中和能源解决方案。

⚡报告中提到的“雄心排放”不包括某些供应链排放，如采购商品、服务和餐饮项目。而谷歌的总排放量实际达到了1518.52万吨二氧化碳当量，相当于近40个燃气发电厂一年的排放量。

Google’s carbon emissions jumped yet again as the company continues to push ahead in AI. The company’s 2025 sustainability report emphasizes that its “ambition-based emissions” grew 11 percent last year to reach 11.5 million metric tons of carbon dioxide pollution, marking a 51 percent increase compared to 2019.

That puts Google farther away from its goal of slashing planet-heating pollution in half by 2030 compared to a 2019 baseline. But what it calls “ambition-based” emissions excludes certain categories of supply chain pollution it considers out of its control or“peripheral” to the core business of Alphabet, Google’s parent company. That includes certain purchased goods and services and food programs.

A table buried in the report’s appendix shows that its total emissions actually reached 15,185,200 metric tons of carbon dioxide. That’s roughly equivalent to the emissions from nearly 40 gas-fired power plants over a year. Google’s report also highlights a 12 percent reduction in carbon emissions from data centers, even though its energy usage continues to rise thanks to AI. Despite the decrease, carbon emissions from data centers are still far higher when compared to 2019.

In its report, Google cites several factors outside its “direct control” that are making it more challenging to achieve its climate goals. That includes the “rapid evolution of AI” that could make the company’s “future energy needs and emissions trajectories more difficult to predict.” It also calls out changes to climate and energy-related policies, “slower-than-needed” deployment of carbon-free energy technologies, and a lack of carbon-free energy solutions in certain markets. “These external factors could affect the cost, feasibility, and timeline of our progress — and navigating them requires flexibility,” Google writes.

AI is driving up energy usage at other companies like Microsoft and Meta, and the technology is on track to consume more power than Bitcoin by the end of this year. Though AI companies like DeepSeek are aiming to create more energy-efficient models, that prospect isn’t stopping companies like Meta from building a massive, gas-powered data center in Louisiana. President Donald Trump has also signed an executive order promoting the  use of coal to power AI data centers.