AI's next bottleneck is power: Chart of the Day
The AI sector is moving from chip shortages to power availability. US data center electricity demand is expected to rise significantly by 2030. This trend is making battery storage a critical part of AI infrastructure.
What changed
New data specifies that US data center electricity demand may rise from 167 terawatt-hours in 2023 to 376 TWh by 2030.
Live updates
-
AI Industry Shifts Focus to Electric Grid Constraints
confidence 90%The AI sector is moving from chip shortages to power availability. US data center electricity demand is expected to rise significantly by 2030. This trend is making battery storage a critical part of AI infrastructure.
What's confirmed:
- The AI industry is shifting its focus from chip shortages to electric grid constraints.
- US data center electricity demand is projected to more than double by 2030.
Still unconfirmed:
- Information transfer speed between chips is emerging as a critical constraint.
-
Power Emerges as Primary Bottleneck for AI Expansion
confidence 90%The AI industry is shifting its focus from chip shortages to electric grid constraints. US data center electricity demand is projected to more than double by 2030. This surge is driving the construction of fast-tracked and off-grid power plants to sustain growth.
What's confirmed:
- US data center electricity demand could more than double by 2030, rising from roughly 167 terawatt-hours in 2023 to about 370 terawatt-hours.
- Power has become a bottleneck in the AI buildout.
- The AI trade has transitioned from a focus on chips to the electric grid.
Still unconfirmed:
- Big Tech is building secret power plants next to AI data centers with government assistance.
- Dozens of off-grid gas plants built for data centers are raising air quality concerns.
- The AI capex cycle is valued at 725 billion dollars.
- AI bottleneck has led to a 300% pricing surge.