In a world undergoing a digital transformation, the energy impact of artificial intelligence (AI) and Bitcoin is generating intense debates. These two technologies, although often viewed through the same lens, have distinct electricity consumption profiles. The challenge is to understand which, between AI and Bitcoin, is truly the champion of electricity consumption, while considering their place in current and future energy systems.
The energy impact of artificial intelligence
The exponential growth of artificial intelligence relies on sophisticated infrastructures, particularly high-performance data centers. According to a study by Selectra, the electricity consumption of these centers could reach around 945 TWh by 2030, equivalent to the annual demand of a country like Japan. This surge in power is primarily linked to generative AI, which requires significant computing capabilities to operate.
The applications of AI, whether linguistic models, image processing, or industrial optimization, generate an increasing demand for electricity. Energy-producing countries, such as France, see this dynamic as an opportunity for modernization. For example, EDF has launched several initiatives to establish data centers near its nuclear and hydroelectric plants, thus driving local consumption with a low carbon footprint.
The energy consumption of Bitcoin
In contrast, the Bitcoin network operates on a decentralized validation protocol based on a proof of work method, which requires constant computing power. The electricity consumption for Bitcoin is estimated at around 188 TWh per year globally by 2025, representing about 0.7% of total consumption. This consumption is relatively stable and concentrated in specific facilities known as mining farms.
These mining farms are often located in regions where energy is both abundant and inexpensive. For example, some countries like Pakistan have integrated cryptocurrency mining into their energy policy, optimizing their surplus electricity production to power these facilities. As a result, Bitcoin’s energy consumption can be integrated more strategically into electrical systems, especially as a complement to renewable energies.
Comparison of energy impacts
When comparing the two technologies, it is crucial to focus on the nature of their electricity consumption. While AI powers a multitude of economic sectors and requires a coordinated response to meet its growing demand, Bitcoin appears as a more isolated player, with specific consumption. In contrast, AI could play a significant role in optimizing energy systems. With efficient algorithms, it can help anticipate demand peaks and manage electricity production more effectively.
The optimization potential of AI
Artificial intelligence not only consumes energy but also offers solutions to optimize electricity production and distribution. Today, projects in Europe are already testing the use of AI to adjust electrical loads, manage maintenance, and enhance the resilience of networks in response to climate variability. Therefore, AI is seen as a crucial technological lever to support the energy transition.
Common challenges: critical resources
Whether for AI or Bitcoin, reliance on specific mineral resources makes them actors to monitor closely from a geopolitical perspective. Indeed, the demand for semiconductors and electronic components is critical for the development of these technologies. The concentration of these resources, particularly in China, poses challenges in securing supply chains.
The goals of autonomy and resilience, for both artificial intelligence and Bitcoin, require significant investment and foresight from governments and industry players. This awareness has prompted the European semiconductor sector to structure a more autonomous supply, aiming to reduce its vulnerability to international tensions.
