Build out of AI data center capacity estimated to reach 117 additional GW by 2030
The chart shows the estimated global AI data center capacity energy demand from 2025 to 2030, broken down by inference and training, alongside power consumption comparisons across select countries. By 2030, inference is expected to account for 69% of AI data center capacity demand.
| Label | Value shown on the chart |
|---|---|
| 2025 Inference | 50% |
| 2025 Training | 50% |
| 2026 Inference | 55% |
| 2026 Training | 45% |
| 2027 Inference | 60% |
| 2027 Training | 40% |
| 2028 Inference | 63% |
| 2028 Training | 37% |
| 2029 Inference | 66% |
| 2029 Training | 34% |
| 2030 Inference | 69% |
| 2030 Training | 31% |
Shown as a preview with credit and a link to the original. Publishers can request removal.
View the original on accel.com
