Budapest hosted the IEEE PES Innovative Smart Grid Technologies Europe (ISGT Europe) conference for the first time between October 19 and 22. IEEE (Institute of Electrical and Electronics Engineers) is the world’s largest technical professional organization with more than half a million members; its Power & Energy Society (IEEE PES) gathers over 40,000 specialists in electricity generation, transmission, distribution, storage and use. ISGT Europe 2026 drew more than 400 participants from 43 countries and featured 284 oral and poster presentations. The event included the presidents of IEEE PES and CIGRE, and it coincided with the 90th anniversary of the Budapest University of Technology and Economics (BME) High Voltage Laboratory.
Current uses of AI and where it may have the biggest impact
The conference theme focused on autonomous, intelligent networks (“From flexibility to autonomy – smart grids for a sustainable and stable future”). István Vokony, head of the Department of Electric Power Engineering at the Budapest University of Technology and Economics, who helped shape the conference program, told our paper that artificial intelligence (AI) already plays an important role in power systems — notably in forecasting, fault detection and supporting operational decisions — and that the set of delegated tasks is expanding.
Vokony said the professional answer to concerns about AI controllability in power systems is gradual, verifiable autonomy: AI can make forecasts and propose interventions, but fully autonomous control requires proven safety limits and reliable fallbacks. This emphasis increases the value of verifiability in research and of testing and clear liability in investments.
Concrete development examples from the conference
Presentations at the event illustrated AI’s practical benefits. C.Y. Chung, president of IEEE PES, showed a model tested on real data from 62 Shenzhen charging stations that forecasts next-day charging demand. Another development combines sky imagery with historical measurements to improve solar PV generation forecasts. Such models can improve storage and consumption scheduling and reduce the costs of unexpected deviations.
An interesting battery example presented was condition monitoring where a large model trained on more than one million charging records transferred knowledge to a small local model; in the test the local model completed fault assessment in 70 milliseconds. That can enable earlier warnings, although researchers emphasized the result does not yet provide a universal safety guarantee for all vehicles.
Cybersecurity: design for resilience and recovery
Cybersecurity was a key topic. Vokony noted that power systems must be designed so that even a successful intrusion cannot cause unlimited damage; this requires segregated systems, strict access controls and practiced recovery procedures. AI can help detect anomalous behavior, but AI systems are themselves attackable. From a supply-security standpoint, prevention must be complemented by the ability to operate during an attack and to recover quickly.
Opportunities and gaps in Hungary
Vokony identified the best opportunities in Hungary for autonomous systems as improving local use of abundant solar generation: better local use can be achieved through storage, smart EV charging, building control and flexible industrial demand. Realizing this requires more measurements, accessible data and market mechanisms that reward flexibility. He said Hungary still lags notably in smart metering and exploiting consumer flexibility.
Strengths include a high share of solar generation and several domestic pilot projects, such as programs by E.ON and the MVM Flexi initiative, which provide a solid base for further development. Nevertheless, Vokony cautioned that Hungary is not yet a general frontrunner in autonomous networks.
Benefits for the domestic knowledge economy and education
Wider adoption of autonomous energy systems could bring high value-added development work, proprietary software and exportable engineering services to Hungary. Achieving this requires specialists who understand both power-system physics and AI methods. According to the conference examples, especially valuable skills will be adapting models to local conditions and validating their outputs; education should therefore build on real operational data, joint industry–university labs and practical assignments.
International cooperation and domestic agreements
Although geopolitical technology restrictions and research-security rules complicate knowledge sharing, significant international collaborations continue. One example mentioned at the conference is the new I-GET research center in Hong Kong, led jointly by Hong Kong Polytechnic University and King’s College London with 60 industry partners.
The event also produced concrete domestic outcomes: the Hungarian Electrotechnical Association (MEE) signed separate cooperation agreements with the Hungarian IEEE PES organization and the CIGRE Hungarian National Committee. These agreements aim to strengthen links between Hungary’s industrial, academic and research communities and to bring international knowledge to local use, with practical value coming from collaborative research tasks and the transition of results into applied solutions.
Conclusion
ISGT Europe 2026 in Budapest underlined that AI is already part of today’s power sector and that in coming years the most visible advances are likely in coordinating PV, batteries, buildings and EV chargers and in optimizing local solar usage. The challenge for Hungary’s research, industry and education is to secure access to data and testing environments and to build training that enables the country to capitalize on these opportunities.
