![](https://cdn.mos.cms.futurecdn.net/KY5Mm3MbVUaApdv4SjXsRN.jpg)
Google's carbon footprint jumped by 48% from 2019, amounting to 14.3 million tons of carbon dioxide emissions for 2023. According to the company's Environment Report 2024, 24% of its total emissions, or over 3.4 million tons, comes from market-based sources, i.e., purchased electricity.
Google plans to achieve net-zero emissions by 2030, and this massive increase looks like a major setback for the company. The company has been actively taking steps to hit this target, but despite its push to use carbon-free electricity sources for its offices and other facilities, data center growth has outmatched the supply of clean energy sources. Furthermore, some areas where Google operates, like in the Asia Pacific, rely more on non-carbon neutral energy sources, meaning the company is forced to purchase 'dirty' electricity while waiting for clean energy projects to come online.
The surge in demand for AI computing has made data centers a popular and growing business. Many firms, including Google, OpenAI, Microsoft, Amazon, Meta, and even Musk's X, are jumping on the bandwagon and procuring millions of GPUs to power their AI dreams. However, we should note that just one of the latest data center GPUs today uses up to 3.7 MWh of power annually. If you multiply this by almost 3.9 million GPUs sold in 2023 alone, that would be enough electricity demand for more than 1.3 million homes.
Industry experts now believe that AI growth will be constrained by the power infrastructure; including power plants, long-distance transmission lines, transformer facilities, and more. With data centers expected to grow by 24.6% year-on-year until 2028, the electricity grid needs some catching up to do, especially as global power production only rose by 2.5% annually during the last 10 years.
However, companies can't just build any power plant to satisfy the electricity demand. Instead, many companies are focused on clean energy sources like wind, solar, and nuclear power plants to avoid increasing their carbon footprint while delivering more power. Microsoft is even investing in research on small modular nuclear plants to power its future data centers. As the globe continues to battle global warming, companies are searching for clean energy sources to power the AI future.