Bitcoin mining and AI computing are both driving massive increases in global electricity demand. While Bitcoin has long been criticized for its energy intensive proof of work system, the rapid scaling of AI GPU clusters is creating a new environmental challenge. Recent data suggests AI emissions are climbing quickly and could surpass those of Bitcoin by 2030. This shift highlights the growing tension between the expansion of high compute technologies and global net zero goals.
Log in to join the discussion
The TWh numbers are fine, but the real issue is where these loads land. Putting a massive AI cluster or a mining farm on a constrained substation without new transmission is a recipe for curtailment. Does the source mention if these loads are moving to areas with stranded renewables?
That is the key. From a policy side, we are seeing more incentives for load shifting. If AI firms can prove they are using curtailed energy, they might get better tax treatment, but the reporting standards for that are still a mess.
Basically a comparison of two huge power sinks. Bitcoin is steady but high, AI is growing exponentially. Worth a read if you want to see the estimated CO2e gap between the two.
Everyone talks about the power draw, but nobody mentions the cooling infrastructure. The water usage for these AI data centers is often a bigger local headache than the carbon footprint. I wonder what the PUE looks like for these new 2026 builds.
True. The carbon numbers are the easy part to track. The actual resource impact on the ground, like water and land for the substations, is where the real friction happens.
