AI Energy Use Surpasses Bitcoin Mining, Creates New Challenges

Artificial Intelligence (AI) is now consuming more energy than Bitcoin mining, sparking concerns and competition over resources. A recent report by the Bitcoin Policy Institute highlights that AI’s energy demands are growing rapidly, with projections showing AI will consume 169 terawatt-hours (TWh) in 2024, surpassing Bitcoin mining’s 160 TWh by 2027, according to Cointelegraph.

While AI’s rise could alleviate criticism directed at Bitcoin mining’s energy usage, it also intensifies competition for electricity and hardware. AI’s profitability—up to 25 times higher per kilowatt-hour than Bitcoin mining—has already led some Bitcoin miners to repurpose their facilities for AI tasks.

AI’s requirement for constant, high availability of power contrasts sharply with Bitcoin mining, which can operate flexibly, using surplus energy and shutting down during grid strain. This flexibility has allowed Bitcoin miners to reduce carbon emissions by 13.6 kilotons in 2023 alone. However, AI’s relentless demand could push miners out of the market, as AI firms with deeper pockets are ready to outbid for power contracts.

The location needs of AI data centers—close to urban hubs—further complicate the competition, whereas Bitcoin mining can thrive in remote areas with abundant renewable energy. Despite the efficiency improvements in AI, the increasing size and power requirements of AI models could negate these gains, much like the advancements in Bitcoin mining hardware led to increased overall energy consumption.

As AI continues to expand, the debate over sustainable energy use will intensify, posing new challenges for both industries.