NextEra, Santee Cooper Report Real AI Savings

The two utilities frame their AI deployments as solving concrete operational and financial problems. NextEra's Grid Composer integrates data from generation, transmission, and trading functions that were previously siloed. Argentieri noted the platform pulls together roughly half a trillion data points a day, and the first version was built in less than 12 weeks. The savings figure of over $20 million this year is attributed to optimized dispatch and outage scheduling, with the tools now offered to other utilities via the Google Cloud marketplace product Optos.

Santee Cooper's motivation centers on weather-forecast errors that can trigger expensive spot-market purchases. Wilson said one degree of temperature variance can cost up to $100,000 per hour on the coldest winter day. The utility serves more than 2 million people in South Carolina. Its financial forecasting group has historically managed over 150 Excel reports, and the new tool is expected to cut runtimes by about 75%. Wilson stressed that AI is not intended to replace jobs, emphasizing a human-in-the-loop approach with a creator-editor approval model for generative AI outputs.

Both utilities have formal governance structures. Santee Cooper is creating an Innovation Council and requires Gemini Enterprise training on technology and confidential data handling. Google Cloud's Smith connected these efficiency gains to Google's data center operations, describing improvements in computing power per unit of electricity. He called the relationship between AI and the grid a virtuous cycle, with efficiency gains feeding into tools utilities can use for power flow modeling.