POTENTIAPower
Energy & Power
Power companies face rising load on constrained grids. We train utility, generation and trading leaders to apply AI to forecasting, dispatch, asset health and interconnection, and to govern it as a critical system.
The landscape
Many power companies are planning for sustained load growth for the first time in a generation. Data centers, electrification and new manufacturing are adding demand, while interconnection queues, transmission limits and equipment lead times slow what can be built. AI sits on both sides of that equation. It drives much of the new load, and it can sharpen how assets are forecast, dispatched, maintained and financed. Arclight trains generation, grid and trading leaders to make those calls with precision, using their own fleet, market rules and regulatory obligations as the material.
- I
Load growth against grid limits
Data centers, electrification and reshored manufacturing are requesting capacity faster than interconnection studies, transmission build-out and equipment supply can deliver it. Every load forecast now carries a capital commitment and a reliability risk.
- II
Reliability with a changing fleet
Retiring thermal units, variable renewables and battery storage change how reserve margins, ramping and outages are planned. The reliability obligation is unchanged, but the fleet is harder to forecast and extreme weather tests it more often.
- III
Every dollar under review
Rate cases, merchant price volatility and investor expectations keep capex and opex under constant scrutiny. Boards must show that AI spending improves reliability, cost or speed to market, and that it is governed like any other critical system.
Where AI earns its place.
Six applications with proven economics in this sector, taught with the techniques behind them and the ways they fail.
- 01
Load, renewable and price forecasting
Forecast load, wind and solar output and nodal prices at the horizons that matter for bidding, hedging and resource planning. Executives learn to read forecast error in dollars and to assign accountability for model performance.
- Probabilistic time-series forecasting
- Gradient-boosted models
- Bayesian hierarchical models
- 02
Asset health and outage prevention
Detect degradation in turbines, transformers, inverters and boilers from SCADA, sensor and inspection data before it becomes a forced outage. The decision is when to intervene, and what that does to availability and the maintenance budget.
- Anomaly detection
- Physics-informed ML
- Survival analysis
- 03
Dispatch, bidding and storage
Optimize unit commitment, battery charging and market bids against physical limits, market rules and price uncertainty. Leaders learn where mathematical optimization already pays, where reinforcement learning is ready and where operators must stay in control.
- Mixed-integer optimization
- Stochastic programming
- Reinforcement learning for control
- 04
Grid planning and congestion
Estimate hosting capacity, congestion and the effect of distributed energy resources across the network, fast enough to test many scenarios. Planners screen options with machine-learned approximations, then confirm the shortlist with full engineering studies.
- Surrogate power-flow models
- Graph neural networks
- 05
Interconnection and site origination
Screen sites against queue position, available capacity, land, permitting, environmental constraints and community factors before committing development capital. Geospatial models turn a long list of parcels into a ranked, defensible pipeline.
- Geospatial ML
- Satellite imagery analysis
- Multi-criteria decision analysis
- 06
Regulatory and contract intelligence
Search and compare tariffs, interconnection agreements, PPAs and regulatory filings with citation-grounded language models. Teams prepare rate case responses, contract reviews and compliance evidence faster, with every answer traced to its source.
- Retrieval-augmented generation
- Document intelligence
- LLM agents
The Energy curriculum.
Taught within the Command Intensive, as an on-site immersion, or privately in one-to-one advisory.
- I
AI across the power value chain
Where AI creates measurable value in generation, networks, retail and trading, and where conventional engineering remains the better tool. Participants map use cases to reliability metrics and the P&L, using their own fleet as the case.
- II
Forecasts, markets and risk
How probabilistic forecasts feed bidding, hedging and resource adequacy. Executives learn to read error distributions, set model risk limits and question a trading model the way they would question a trader.
- III
Operational technology and safety
Deploying models near SCADA, energy management and control systems without compromising safety or NERC CIP compliance. Covers human-in-the-loop design, cyber exposure and change control for models that influence physical equipment.
- IV
Capital allocation and governance
Managing AI as a capital program, with business cases, stage gates, vendor selection and build-or-buy decisions. The module closes with a board-ready plan the team can defend to regulators and investors.
Who attends
- CEOs and executive committees
- CFOs and heads of corporate development
- COOs and fleet operations leaders
- Heads of trading, origination and risk
- Transmission and grid planning directors
- Chief digital, data and technology officers
Outcomes
- A ranked AI portfolio tied to reliability, margin and capital plans
- The fluency to challenge forecasts, models and vendor claims
- Governance suited to operational technology and reliability standards
- A 12-month roadmap with owners, stage gates and value measures
How energy leaders engage us.
- IThe Command IntensiveIMPERIVMThree days to set the AI agenda for the enterprise.3 daysUp to 12 executives
- IIOne-to-One AdvisoryCONSILIVMPrivate AI counsel, built around your real decisions.90 daysOne executive
- IIIOn-Site ImmersionsIN SITVExecutive training at your assets, built on your data.2–5 daysLeadership teams up to 30
- VSector IntelligenceSPECVLADecision-grade analysis of AI in your sector.From 72 hoursExecutive teams
Begin
A sector briefing for your leadership, within seventy-two hours.
Tell us about your assets and the decision in front of you. A principal replies within one business day.