A demonstration project achieved a 27% peak reduction and 12.5% bill savings using distributed AI
Providence, RI, May 15, 2025 (GLOBE NEWSWIRE) — Today, AI technology company Utilidata announced the successful completion of a demonstration project in collaboration with EPRI’s Incubatenergy® Labs (IEL) and Southern California Edison (SCE). The project showed the capabilities of Utilidata’s Karman AI platform in delivering real-time, autonomous management of distributed energy resources (DERs) from the electric meter. Utilidata’s technology proved that local dynamic DER management is possible by enabling real-time operational signals to DERs – overriding static dispatch schedules. This allows DERs to actively respond to real-time system needs rather than pre-scheduled events.
Using wireless communication and on-device AI, Karman calculated optimal DER dispatch and sent control signals directly to DERs – while keeping data local at the source (the meter). This approach will enhance resiliency, reduce data transmission costs, maintain customer privacy, and increase customer benefits. The demonstration project resulted in a 12.5% reduction in electricity cost for a simulated home and a 27% reduction in peak demand.
Utilidata was selected by SCE to demonstrate the Karman platform for load forecasting, solar disaggregation and forecasting, local DER optimization calculations and DER commands. The study was performed using EPRI’s Simulation Platform for Integration of DER (SPIDER) software which generated realistic net load data.
“Delivering scalable, real-world solutions for the future grid, such as distributed energy resources, requires deep collaboration between utilities and technology innovators, and EPRI’s Incubatenergy Labs facilitates these partnerships,” said Dan Killoren, Head of the Global Innovation Hub for EPRI. “This project demonstrates that AI could be used to enhance grid operations.”
In addition to accurately identifying multiple DERs in real-time, during the simulation, the Karman platform demonstrated the following advanced capabilities:
- AI-driven load and solar forecasting: With real-time insight into local conditions, Karman enabled accurate forecasting of site load and solar output. In the field, Karman can allow for more dynamic rate schedules.
- Generating and wirelessly sending operational signals to DERs: Unlike traditional systems that rely on centralized, static dispatch, Karman autonomously created and sent operational signals to DERs – curtailing solar generation and commanding batteries to charge or discharge – all without sending data to the cloud. This allows DERs to actively respond to real-time system needs, rather than pre-scheduled events.
- Achieve peak reduction and cost savings for customers: By aligning DER control with time-of-use rates and peak pricing periods, Karman reduced peak load and lowered energy costs for the end user – demonstrating how AI at the grid edge benefits both utilities and consumers.
“This project is the first to show how distributed AI embedded in local grid devices can forecast, manage, and optimize DERs in real time,” said Jess Melanson, President & Chief Operating Officer at Utilidata. ”To realize the full potential of DERs, utilities need the ability to rapidly and dynamically respond to system needs using local computing power.”
Current DERMS and virtual power plant (VPP) systems often rely on pre-determined schedules and lack awareness of local grid constraints. Utilidata’s Karman platform changes that paradigm by embedding AI directly into grid edge devices, creating a responsive, intelligent layer of local control. It enables DERs to act as dynamic participants in the grid – not passive assets awaiting pre-scheduled events.
SCE has become a leader in accelerating the energy transition by pursuing innovative technologies within grid edge computing to address growing energy demand and grid complexities. SCE’s recent partnership with NVIDIA aims to bring more AI-based solutions to the grid, and its leadership in the industry consortium, the Grid Edge Interoperability & Security Alliance (GEISA) project, has established important industry collaboration to set specifications for deploying grid edge solutions in the utility industry.
Built on a custom NVIDIA module designed specifically for grid operations, Karman’s distributed AI platform delivers secure over-the-air updates and an open architecture for third-party applications. DERMS and VPP providers can build applications directly on the Karman platform, expanding the ecosystem of intelligent DER management tools.
About EPRI:
Founded in 1972, EPRI is the world’s preeminent independent, non-profit energy research and development organization, with offices around the world. EPRI’s trusted experts collaborate with more than 450 companies in 45 countries, driving innovation to ensure the public has clean, safe, reliable, affordable, and equitable access to electricity across the globe. Together…shaping the future of energy.®
About Utilidata:
Utilidata is an AI-powered technology company providing utilities with real-time, actionable insights to operate a dynamic and flexible grid. The company’s distributed AI platform, Karman, analyzes large volumes of electricity data in real-time leveraging a combination of on-chip and cloud based software. Powered by a custom NVIDIA module purpose-built for the grid, Karman provides industry-leading computational power to grid devices–starting with meters–enabling faster, smarter decision-making.
CONTACT: Jessica Ridlen Utilidata jridlen@utilidata.com
- DV DEADLINE NOTICE: ROSEN, LEADING INVESTOR COUNSEL, Encourages DoubleVerify Holdings, Inc. Investors to Secure Counsel Before Important July 21 Deadline in Securities Class Action – DV - July 19, 2025
- Arctic Pablo Coin Launches Weekly Myth-Themed Presale Structure with $0.0005 Entry Price and Token Burn Mechanism - July 19, 2025
- RCAT DEADLINE ALERT: ROSEN, A HIGHLY RECOGNIZED LAW FIRM, Encourages Red Cat Holdings, Inc. Investors to Secure Counsel Before Important July 22 Deadline in Securities Class Action – RCAT - July 19, 2025
