Summary
- National Grid and Keen AI have secured £355,985 in Alpha phase Strategic Innovation Fund backing for FoSMo.
- The project brings together Britain’s three transmission operators and selected distribution operators to pool anonymised visual asset data.
- Better grid asset monitoring supports reliability and expansion at a time when large load connections are harder to deliver.
National Grid and Keen AI have secured £355,985 in Alpha phase funding to develop a shared AI model for electricity network asset monitoring across Great Britain.
The project, Foundational Shared Model Operations, or FoSMo, is backed by the Strategic Innovation Fund, an Ofgem programme managed with Innovate UK. It is designed to standardise how network operators collect, analyse, and act on visual data from assets such as pylons, cables, insulators, and fittings.
The Alpha phase brings together National Grid Electricity Transmission, SP Energy Networks, and SSEN Transmission, covering all three electricity transmission operators in Great Britain. Distribution operators UK Power Networks and Electricity North West are also involved.
Operators want a larger inspection dataset
Electricity networks operate as connected physical systems, but AI tools for monitoring assets are often developed separately by individual operators. Separate development can create duplicated work, smaller datasets, and weaker performance when rare defects appear too infrequently inside one operator’s asset base.
FoSMo is intended to pool anonymised visual data from participating networks and create a foundational computer vision model that operators can adapt for their own needs. Keen AI, which says it has processed more than one billion images for UK electricity transmission and distribution customers, will develop, maintain, and host the models in the UK.
National Grid says full adoption across participating operators could save the industry about £22.6m over five years from 2027. It also projects that improved asset monitoring could prevent around 85,000 consumer interruptions and 5.2m minutes of power loss each year once fully adopted.
The project sits upstream of data centre delivery but close to a major sector constraint. Large data centre campuses depend on reliable transmission and distribution networks, and the UK is already managing a crowded grid queue containing renewables, storage, industry, housing, and digital infrastructure.
Grid growth needs better asset data
Building more substations, circuits, and connections will not be enough if operators cannot inspect and maintain a larger asset base efficiently. Grid expansion also needs better asset data, defect detection, maintenance prioritisation, and governance models for shared information.
The project also has a domestic capability element. A UK developed model, using data owned by the industry and hosted in the UK, may reduce some governance concerns associated with critical infrastructure data being processed through fragmented third party systems.
FoSMo will not immediately shorten a data centre connection queue or deliver a new substation. Its value is more systemic: better monitoring can reduce fault risk, improve maintenance decisions, and support reliability as operators build and operate more assets under heavier demand.
The Strategic Innovation Fund’s wider Cycle 5 portfolio also includes projects examining flexible demand and large load integration. Those projects reflect a practical issue for the UK power system, where new grid capacity is arriving more slowly than demand from some large users.
FoSMo remains an Alpha phase project, so it still has to prove its model design, data governance, accuracy, and adoption across multiple operators. The underlying need is clear enough: if the UK wants to connect more large loads while electrifying the wider economy, the grid will need better visibility over the assets already in service.
