LF Energy adds four grid projects as utilities share more of the plumbing

Coreso and Landis+Gyr joined as members while projects spanning asset planning, grid AI and home energy entered LF Energy's portfolio.

By · Published

Primary source: Linux Foundation Newsroom

Why it matters

LF Energy is turning utility-built software into shared infrastructure, reducing duplicated engineering while giving vendors a common layer on which to compete.

A modern street scene at dusk, featuring a sleek utility pole, a smart meter on a building, and subtle light patterns representing energy flow.

Alex Thornton expanded LF Energy on September 15th with two new members and four open source projects, pushing founder Shuli Goodman's utility collaboration model deeper into grid AI, asset planning and home energy management.

The additions, announced during LF Energy Summit Europe in Berlin, include European grid coordination center Coreso and smart-meter manufacturer Landis+Gyr as general members. LF Energy also accepted AssetLife, CityLearn, EnerGNN and Smart HEMS Benchmark, while several existing projects advanced through LF Energy's technical maturity system.

Thornton inherited an organization built through Goodman's persistence. In the mid-2010s, Goodman drove to the Linux Foundation's San Francisco office and arrived without a scheduled meeting to argue that utilities needed the kind of shared software development already common in cloud computing and telecommunications. Her conversations with French transmission operator RTE and other institutions led to LF Energy's public launch in July 2018.

Goodman, who died from cancer on January 3rd, 2023, saw open source governance as a way for regulated utilities to share foundational technology without handing control to a single vendor. Thornton, a former CTO of carbon and energy-management software developer Cleartrace, took over as executive director in December 2023 after earlier roles at Locus Energy and Sunrun.

The September 15th expansion follows the same thesis: utilities can contribute the software they already use, vendors can build commercial products above it, and LF Energy can handle the governance and security work that makes collaboration acceptable to legal departments and regulators.

Four projects at four different points

The new projects cover a wide technical range, though their maturity and deployment evidence vary.

AssetLife is a Python library originally developed and used in production by RTE. It applies survival analysis, stochastic modeling and renewal theory to infrastructure records, helping asset managers compare the expected costs of preventive replacement, repairs and running equipment until it fails. LF Energy lists RTE, TenneT, AusNet and gas transmission operator GRTgaz as production users.

That is a practical use of open source for a problem utilities usually solve with internal spreadsheets or proprietary planning tools. AssetLife can also include a carbon shadow price in its calculations, allowing an operator to weigh societal costs alongside direct maintenance and replacement spending.

CityLearn is a simulation environment for testing how batteries, heat pumps, electric vehicles and solar installations can respond to conditions across a district. Researchers can compare rule-based systems, model-predictive control and reinforcement learning against the same datasets and performance measures. Contributors include the University of Texas at Austin's Intelligent Environments Lab, TU Eindhoven, Politecnico di Torino and Concordia University.

EnerGNN applies graph neural networks to power systems. Built with JAX and Flax, EnerGNN models the varied components and connections found in real transmission networks instead of reducing a grid to a simple set of interchangeable nodes and edges. RTE is using EnerGNN in research and development work involving French transmission data with more than 7,000 buses.

RTE's tests illustrate both the attraction and the limit of the approach. In one topology optimization experiment, EnerGNN produced a result in roughly 200 milliseconds, compared with about 10 minutes for a mixed-integer linear programming baseline. The faster result delivered a smaller capacity improvement, and LF Energy describes the work as an R&D demonstration rather than control-room deployment.

Smart HEMS Benchmark is intended to give researchers, manufacturers and utilities a shared method for comparing home energy-management systems. It covers equipment sizing, daily household dispatch and virtual-power-plant support using common datasets, scenarios and metrics. Its open source launch remains planned for late 2026.

All four entered LF Energy at the Sandbox stage, the entry point for projects that warrant experimentation and need to establish governance, security practices and a broader contributor community. The distinction matters because Thornton described LF Energy's growing collection as delivering "production-ready open source tools." AssetLife has named production users, while EnerGNN remains in R&D and Smart HEMS Benchmark has yet to complete its planned public release.

Membership pays for the neutral layer

LF Energy is a Linux Foundation initiative supported through memberships and contributed engineering resources. Its published membership schedule prices general membership from $5,000 to $70,000 annually, depending on an organization's size and whether it already belongs to the Linux Foundation.

Members receive staff support, industry introductions, visibility and a path to governing-board candidacy. That arrangement gives Coreso and Landis+Gyr a stake in the shared layer beneath commercial energy products. Coreso coordinates electricity flows across European borders, where compatible grid models and operational tools carry direct value. Landis+Gyr sells meters and energy-management technology to utilities, giving it an incentive to reduce integration friction while competing on the products built above the common code.

LF Energy says it now hosts more than 35 projects supported by over 80 member organizations and 270 contributing organizations. The organization also reported that summit registration increased by more than 30% from the previous year. Registration growth demonstrates interest in the community; deployments, independent contributors and sustained maintenance will provide the harder measures of whether that interest becomes durable infrastructure.

Existing projects move up

The stronger evidence of maturity sits elsewhere in LF Energy's portfolio. SEAPATH, a virtualization platform for digital substations, advanced to Graduated status, which LF Energy reserves for projects with established production use, mature security practices and contributions from multiple organizations. CoMPAS moved to Early Adoption, while Grid2Op and OpenGridFM advanced from Sandbox to Incubation.

LF Energy also released OpenSTEF 4.0, a ground-up restructuring of its short-term energy forecasting software. The release separates the system into modular packages for core data types, forecasting models, model selection, backtesting and pretrained foundation-model forecasters. LF Energy says Dutch grid operator Alliander runs OpenSTEF at thousands of grid locations and Swedish consultancy Sigholm uses it to optimize about 40% of Sweden's district-heating production.

Thornton said open source is becoming "a core strategy for digital energy infrastructure." His task is to make that strategy survive the slow procurement cycles, security requirements and institutional caution of the power industry.

The latest intake gives LF Energy more experiments to govern and more organizations willing to fund the work. Goodman's original insight still sets the standard for judging the result: shared code becomes infrastructure only when competing institutions trust it enough to maintain and operate together.

Reader comments

Conversation for this story loads after sign-in.