The rapid expansion of artificial intelligence is increasingly constrained by electricity grid limitations, forcing tech companies to secure dedicated power sources to meet strict infrastructure deployment timelines.

Key Points

• Global grid connection queues currently hold over 2,500 gigawatts of projects, including significant data center demand.
• Texas grid operator ERCOT reports 438 GW in interconnection requests, with nearly 90 percent attributed to data centers.
• Major tech firms like Microsoft, Google, and Amazon are bypassing public grids by investing directly in nuclear and gas-fired power projects.
• FERC has issued June 2026 orders requiring grid operators to justify how they manage large-load connections and co-location arrangements.
• Sovereign wealth funds from the Gulf are increasingly providing direct equity to bridge financing gaps for energy-intensive AI infrastructure projects.

Why it Matters

The ability to deliver reliable, dispatchable power within a 24-to-48-month window has become the primary determinant for where AI infrastructure is built. Jurisdictions that fail to integrate energy planning with compute requirements risk losing investment to regions that offer regulatory fast lanes and sovereign-backed capital.