Case Study: National Grid achieves 40% more accurate renewable-energy forecasts with Oracle Cloud Infrastructure

National Grid analyzes weather 40% better on Oracle Cloud

National Grid ESO, which manages Great Britain’s electricity system, faced the challenge of integrating vastly more variable renewable generation—millions of small solar panels and wind turbines—into the grid. To hit carbon-reduction targets and reliably match supply and demand, the team needed highly accurate, fast machine‑learning forecasts and short bursts of powerful compute that their prior setup couldn’t provide.

By moving to Oracle Cloud Infrastructure (including Oracle Cloud Compute) and building a virtual supercomputer (“Mildred”) with NVIDIA GPU‑based cores, National Grid ESO ran tens of thousands of models and cut runtimes from hours to minutes. Model accuracy improved by about 40%, the team uncovered previously unseen patterns, and the shift freed them to focus on strategy—contributing to Great Britain reaching 48.5% electricity from renewables for the 12 months ending December 2019.

National Grid

James Kelloway

Energy Intelligence Manager