Artificial intelligence is driving a surge in data center electricity demand that results in a third of that power being wasted [1].
This inefficiency presents a critical challenge for the energy economy as the rapid expansion of AI infrastructure threatens to outpace the development of new power generation sources.
Katie McGinty said the energy economy’s biggest waste problem is already inside the system. The scale of the issue is tied to the accelerating growth of computing needs required to sustain large-scale AI models.
According to McGinty, data center demand is projected to reach 945 TWh by 2030 [2]. She said the fastest new energy source available is not the creation of more power, but rather the recovery of the third of electricity currently wasted [1].
Data centers require immense amounts of power for both processing and cooling. When this energy is not recovered or used efficiently, it becomes a systemic loss that increases the overall burden on the electrical grid, a problem that scales alongside AI adoption.
"The fastest new energy source isn't generation — it's the third of electricity quietly wasted," McGinty said [2].
“"The energy economy’s biggest waste problem is already inside the system"”
The reliance on AI is creating a paradox where the technology intended to optimize systems is stressing the physical energy grid. By identifying waste recovery as a primary 'energy source,' the focus shifts from building new power plants to improving the thermal and electrical efficiency of existing infrastructure to prevent a systemic energy crisis by 2030.
