Utilities are using AI to manage rising power demand from AI workloads. Current deployments focus on demand forecasting and maintenance prioritization. Uneven AI outputs and lack of auditability limit broader use. Most utilities cannot reconstruct AI system decisions. Regulators in New York and Connecticut are increasing scrutiny of utility AI use. A survey found 70% of energy firms see AI as a cybersecurity risk. Only 23% of utilities have a process to manage that risk. Nearly 80% of utility leaders have deployed at least one AI application. Outdated frameworks and cost shifts to residential customers hinder AI adoption. Examples include identifying 80 MW of grid capacity early and reducing data center power demand by up to 40%.
What changed
Utilities relied on traditional forecasting and manual maintenance prioritization without AI support.
What it unlocks
Using AI for demand forecasting, maintenance prioritization, and identifying available grid capacity earlier.
- 70% of energy and utility companies see AI as a cybersecurity risk
- 23% of utilities have a process for managing AI risk
- 9% of energy companies have a process for managing AI risk
- 80% of utility leaders deployed at least one AI application
What you need to act on it
- high-quality input data
- governed processes for auditability
- domain expertise combined with AI fluency
Sources