GLOBAL RESEARCH ARCHIVE
North American Power & Utilities: NARUC 2026 Summer Policy Summit Takeaways and Investment Implications
Research evidence excerpt
North American Power & Utilities: NARUC 2026 Summer Policy Summit Takeaways and Investment Implications
demand and toward whether the sector can provide enough transparency to convert that
demand into approved, recoverable, and publicly acceptable investment.
The clearest through-line across the panels was that data centers are no longer being
treated as routine large-load additions. They are now system-planning events. AI-driven
load is large, concentrated, fast-moving, and highly location-sensitive, which means it
can affect generation needs, transmission planning, distribution infrastructure, gas
supply, water resources, local permitting, and customer bills at the same time. Panelists
pointed to the scale of recent development activity, including roughly $120bn of data
center investment cited for Virginia, as evidence of the opportunity. However, they also
made clear that a large pipeline is not the same thing as a financeable or
interconnectable pipeline. Regulators are increasingly looking for proof that projects are
real, durable, and ready to move through the queue before allowing utilities to commit
capital that may ultimately be recovered from customers.
For investors, this distinction matters because the investment case is becoming more
selective. Utilities with credible large-load processes, clear tariff structures, timely
recovery mechanisms, constructive commission engagement, and visibility into customer
commitments should be better positioned to translate data center growth into approved
capital investment and earnings visibility. Conversely, utilities that rely on speculative
load forecasts, opaque interconnection queues, or weak cost-causation protections may
face higher disallowance risk, political pushback, or project delays. The panels therefore
The English excerpt is extracted automatically from the cited source page and may contain layout or recognition errors. It is never batch translated.
Open report viewer