Caleb Harding and Lily Ottinger
AI development is facing an imminent electricity bottleneck. Data centers collectively make up 3% of total US electricity demand, and some predictions indicate they could consume up to 8% of electricity demand by 2030.1 And already, energy-consumption concerns have stalled some US data-center construction.
This leaves companies in a tough spot: if they can’t get enough energy to train their models, they will fall behind.
In response, some companies have quietly walked back promises of carbon neutrality, pivoting away from renewables and burning fossil fuels to pick up the slack. Meanwhile, Amazon has purchased a data center right next to a nuclear power plant to ensure an adequate supply of green energy.
China is staring down the same bottleneck. Today, we’re exploring China’s plan to unlock green energy abundance on the journey to become a world leader in AI.
No comments:
Post a Comment