Space weather events, such as solar storms, can drive geomagnetically induced currents (GICs) in power grids, pipelines, and undersea cables. These currents may damage transformers, disrupt electricity supply, and threaten other critical infrastructure. Existing global indices like Kp provide only coarse warnings and do not reflect local vulnerabilities, leaving operators without actionable insight. Our project is buildnig a deep learning–based forecasting model that integrates solar wind data, ground magnetometer measurements, and historical GIC events to produce probabilistic, location-specific forecasts. Tests on past storms show that the model outperforms existing tools. The system can issue “watch” and “warning” levels, allowing operators to take preventive measures such as load redistribution or temporarily disconnecting equipment. While initially focused on the Norwegian power grid, the approach is adaptable to other infrastructures, including pipelines, railways, and communication cables, and is attracting interest from sectors such as insurance and space operations. By improving preparedness and resilience, this technology addresses a critical gap between awareness of space weather risks and the ability to mitigate them, supporting the transition to sustainable, reliable, and electrified societies.
The qualification project has delivered both technical and commercial advances towards a novel forecasting tool for geomagnetically induced currents (GICs). On the technical side, we have moved from a deterministic proof of concept to a deep learning–based probabilistic model that ingests real-time solar wind and geomagnetic measurements to provide location-specific forecasts. Tests on past storm events confirm that the model can reliably capture short-term fluctuations and provide useful guidance several hours in advance, outperforming existing global indices. This demonstrates that the concept is feasible and raises the technology readiness level to TRL4. The model’s ability to issue probabilistic “watch” and “warning” thresholds offers operators actionable lead times to mitigate risks by redistributing loads or temporarily disconnecting vulnerable equipment.
On the commercial side, the project has clarified the potential market fit of such a tool. Outreach to more than 60 companies across multiple sectors highlighted both challenges and opportunities. The broader societal impact lies in improving resilience to space weather at a time of accelerating electrification. By reducing the probability of blackouts and costly equipment failures, the tool supports critical infrastructure reliability and energy security. It also aligns with several UN Sustainable Development Goals, including affordable and clean energy, resilient infrastructure, and sustainable cities. Economically, even modest improvements in preparedness can prevent multimillion-dollar damages and contribute to long-term grid stability, thus reducing systemic risks for utilities, insurers, and society at large.
The potential impacts extend internationally. As global solar activity increases towards the 2025 maximum, demand for forecasting solutions is expected to grow, and the project positions Norway at the forefront of this emerging market. With further development and testing, the technology could become a key element of critical infrastructure protection, contributing both to national preparedness and to a wider international market for space weather resilience.
