Member-only story
Artificial Intelligence and the use of power
A boon to resources or a drain?
In October of last year, Google signed a deal to buy power from six or seven small modular nuclear reactors. The move will see Kairos Power, a California based company, founded in 2016, build nuclear reactors based on their novel, advanced reactor technology.
The deal underscores the increasing need for the ultra-high, energy-intensive AI industry. A single generative AI query requires nearly 10 times more electricity than a traditional internet search.
Large scale AI developments lead to increased energy levels, which not only puts pressure on the existing grid but also raises the operational costs for companies. Researchers recently found that the cost of the computational power required to train these models is doubling every nine months. This increase in energy demands will also favour areas where energy costs will be at the lower end of the scale, possibly disadvantaging areas with limited supply. This will inevitably lead to higher costs and will influence pricing and therefore potential accessibility for AI driven products and services.
According to the International Energy Agency (IEA) electricity usage by data centres will increase significantly in the near future thanks to the demands of AI and cryptocurrency. In the United States, approximately 4%…