ADI built a bottom‑up view of U.S. data center electricity demand driven by AI workloads, using rack density, GPU power consumption, utilization assumptions, and load growth scenarios. These demand estimates were stress‑tested through interviews with utilities, hyperscalers, and power developers to assess grid constraints and timing mismatches. The work linked physical load growth to regional power pricing outcomes over the next decade.
The client
Integrated energy company
The situation
Rapid AI growth created uncertainty around load growth, grid constraints, and power pricing.
ADI’s contributions
Proprietary demand modeling
Modeled load growth using bottom‑up GPU and rack‑density assumptions.
Primary market interviews
Interviewed utilities, data center operators, and OEMs.
Supply constraint analysis
Assessed transmission and grid bottlenecks by region.
Pricing outlook
Developed electricity price trajectories across scenarios.
Key outcomes
- Established demand and price scenarios to support gas‑to‑power and infrastructure strategy.
More insights
SAF Tracker
2026 June SAF Tracker highlights – #88
Here are the latest highlights from ADI’s SAF Tracker: The full newsletter along with archives and databases are available to SAF Tracker subscribers.
News
Utility capital projects & 2026 energy trends
Utility capital projects are facing mounting delays and cost pressures in 2026, even as AI-driven demand fuels record capex. Meanwhile, the upstream oil and gas market is stabilizing around a normalizing shale cost curve, bulk liquid storage operators are shifting toward capability-driven growth, and the energy transition is exposing critical minerals bottlenecks that are pushing […]
Article
U.S. refining capacity is gradually consolidating into larger, more complex facilities
U.S. refining capacity shows limited overall growth, but the structure of the system is shifting. Expansions at large, complex refineries are driving changes on the supply side, while smaller plants face cost and operational constraints that are forcing exits. This is steadily concentrating capacity in fewer, more sophisticated facilities. Key drivers capacity consolidation: Geographic concentration […]