Bright key facts / Emerging
AI data centers and the grid bottleneck: what real progress would look like
A new Goldman Sachs forecast highlights the scale of the buildout. The more useful test is whether new computing demand brings reliable electricity, workable connections and fair terms for the people sharing the grid.
- AI’s role
- AI computing creates electricity demand while efficiency per useful task and flexible operation can influence the infrastructure needed.
- Documented result
- Goldman Sachs’s October 19 analysis forecasts global data-center power demand increasing 170% from 2025 to 2030. The IEA separately projects 485 TWh in 2025 rising to about 950 TWh in 2030. Berkeley Lab counts 1,312 GW of generation and approximately 749 GW of storage seeking transmission connections at end-2025.
- Important limitation
- Forecasts are source-attributed projections, not measured AI consumption or guaranteed construction.