Everybody in the Pool E129: The AI Solution the Grid Desperately Needs with Gridmatic
Apr 2, 2026
David Miller, Chief Commercial Officer at Gridmatic, builds AI that forecasts renewables and optimizes batteries. He discusses why renewables are harder to manage than fossil plants. He explains day-ahead prediction of price spikes, using batteries and flexible loads, and how controllable load resources and virtual power plants can unlock faster interconnection and reliable clean power.
AI Snips
Chapters
Transcript
Episode notes
AI Replaces Traditional Grid Modeling For Renewables
- Gridmatic uses AI to manage renewable-backed retail electricity for commercial customers.
- Founder repurposed ad-targeting ML to predict where and when wind will blow and when the grid will be stressed.
Renewables Are Much Harder To Forecast Than Demand
- Renewable output is far more volatile than load so relying on renewables requires better forecasting and risk management.
- Wind hourly forecast errors can be ~15% versus load errors of only a few percent, driving need for advanced models.
Statistical AI Beats Trying To Replicate The Whole Grid
- Gridmatic avoids full power-flow replication and instead trains statistical AI on abundant grid outputs and weather data.
- They combine topology, historical production, and custom weather models to improve day-ahead predictions.