CryptoRank warns AI may soon outstrip Bitcoin in global energy use

This report crystallizes an emerging reality: the bottleneck for frontier AI is shifting from chips and capital to electrons and grid capacity. While Bitcoin spent years as the poster child for digital energy use, CryptoRank’s analysis shows AI data centers already consuming roughly twice Bitcoin’s electricity and accelerating. Training and serving ever‑larger models, plus embedding them into search, productivity tools and embedded devices, is turning AI into a quasi‑baseload load on power systems around the world.

For the AGI race, that means timelines are increasingly constrained by concrete infrastructure—power plants, transmission lines, water for cooling—rather than just algorithmic cleverness. In the short run, this tends to accelerate progress because operators race to secure scarce interconnects and sign aggressive power purchase agreements. In the medium term, however, political and environmental pushback on data center siting and emissions could slow or redirect growth. The comparison with Bitcoin is telling: AI has so far avoided the same level of scrutiny because its consumer value is more obvious, but once it claims a percentage point or more of global demand, it will take center stage in climate and energy debates.