Hong Kong's data centers are among the world's most carbon-intensive, according to a recent UN study, highlighting the city's heavy reliance on fossil fuels. This issue is particularly concerning given the rapid growth of AI, which is expected to consume vast amounts of energy by 2030. The study, titled "Environmental Cost of AI's Energy Use," reveals that data centers could consume 945 terawatt-hours by 2030, nearly tripling the combined annual electricity use of Pakistan, Bangladesh, and Nigeria. This is a stark reminder of the environmental impact of our increasing reliance on technology.
What makes Hong Kong's situation unique is its heavy dependence on fossil fuels, with 67% of its energy derived from fossil fuels, compared to the global average of 43%. This high reliance on fossil fuels contributes significantly to the city's carbon footprint, which is 43% higher than the global average. In contrast, countries like the US, Germany, and Italy have carbon footprints 18%, 24%, and 32% below the global average, respectively, thanks to their greater use of renewable energy sources.
The environmental impact of data centers extends beyond carbon emissions. They also require substantial water and land resources for cooling and infrastructure. However, Hong Kong's energy mix, which lacks hydropower and renewable energy sources, results in lower water and land consumption compared to other regions. Despite this, the city's status as a trade and logistics hub, coupled with its extensive telecommunication networks, makes it an attractive location for data centers.
One of the most intriguing aspects of the study is the distinction between the energy demands of AI models during daily use and model training. The report reveals that daily use of AI models accounts for around 80-90% of total energy demand, with image generation requiring a thousand times more energy than a simple text query. This highlights the need for more efficient AI models and practices to reduce the environmental impact of AI.
To address these concerns, the Hong Kong government is taking steps to build a new 110,000 square meter data facility in Sandy Ridge, with 90% dedicated to data centers. Additionally, the Electrical and Mechanical Services Department (EMSD) has implemented policies to ensure energy efficiency in buildings and regulate the use of fresh water in cooling towers. The EMSD's Green Data Centers Practice Guide promotes efficient data center design, procurement, operations, and disposal, while also encouraging the use of assessment tools to measure environmental impacts.
In conclusion, Hong Kong's data centers are among the world's most carbon-intensive, primarily due to the city's heavy reliance on fossil fuels. This issue is exacerbated by the rapid growth of AI, which is expected to consume vast amounts of energy. However, the city's government and data center operators are taking steps to address these concerns, such as building new data facilities and implementing energy efficiency policies. It remains to be seen whether these efforts will be sufficient to mitigate the environmental impact of Hong Kong's data centers.
