Artificial intelligence could surpass Bitcoin in energy consumption as early as the end of 2025, according to an analysis by Alex de Vries from the Amsterdam University of Applied Sciences.
According to his forecasts, AI will account for about 23 GW, comparable to the energy consumption of the United Kingdom. This is about half of the total consumed by all data centers in the world.
De Vries, who previously tracked the impact of cryptocurrencies on energy, published data in the journal Joule. He used the 'triangulation' method, comparing the production of TSMC chips, company data, and reporting to estimate the growth of energy costs associated with AI. In just one year, the output of chip packaging for this sector has doubled.
The scientist notes the similarities between crypto mining and the AI sector. The rapid growth of data centers, lack of transparency in reporting, and increasing load on the electricity grid. Despite promises of climate responsibility, companies like Microsoft and Google do not disclose data separately regarding AI in their sustainability reports.
To meet demand, energy companies in the USA have already announced plans to build gas and nuclear power plants. This creates risks for the transition to green energy, similar to what is happening with mining. The problem is compounded by the fact that the efficiency of AI is increasing, but its usage is growing even faster.
The question of whether the increase in energy efficiency will lead to a reduction in overall energy consumption remains open. Experts point to the risk of the Jevons Paradox: the more efficient AI becomes, the more it is used. Without detailed and transparent data on energy consumption in this sector, solving the problem is impossible, De Vries emphasized.
#CEXvsDEX101 #TradingTypes101 #TrumpMediaBitcoinTreasury #PCEMarketWatch #MarketPullback
