Why American data centers can't plug in
The article examines why new data centers in the US face long delays connecting to the electric grid, citing broken interconnection processes, regional grid constraints, and regulatory bottlenecks that hinder the rapid deployment of power-hungry AI and cloud computing infrastructure.
Background
- For decades, data centers could reliably get grid interconnection in 1–2 years; that timeline has now ballooned to 3–7+ years in much of the US, and some projects face 10+ years or outright rejection.
- The root cause is not a shortage of electricity generation, but a bottleneck in the **interconnection queue** — the process by which new facilities request permission to connect to the high-voltage transmission grid.
- This queue is managed by **Regional Transmission Organizations (RTOs)** and **Independent System Operators (ISOs)**, such as PJM, MISO, and CAISO. These non-profit entities run the grid across multiple states. But they lack the staff and software to process the surging volume of requests.
- The surge is driven by three converging trends: the AI boom (enormous compute clusters), the buildout of renewable and battery projects (many more small generators), and the slow retirement of coal and nuclear plants, which complicates grid stability.
- Thus, a technical, regulatory, and administrative pile-up — not a lack of power — is the real constraint on American data-center growth and AI infrastructure.