At the Portfolio AI in Energy 2026 conference on May 27, 2026, industry participants made clear that artificial intelligence is shifting from pilot projects and presentations into everyday operational use across the power sector. Sessions examined how corporate data can be converted into measurable financial value, how algorithms are integrated into decision-making, and where investments generate true competitive advantage versus costly developments with poor returns.
Fast prototyping and short test cycles
Several speakers noted that current technologies and digital tools allow rapid prototyping and real‑world testing of new ideas. Veszprémi Balázs, CEO of ARTEMIS, said that companies that do not innovate risk losing their market position, since legacy methods increasingly fail; optimization now requires second‑level re‑planning rather than hourly or daily decisions.
Measurable value and the need for KPIs
Varga Pál, head of the Department of Telecommunications and Artificial Intelligence at the Budapest University of Technology and Economics (BME), stressed that the success of AI projects is determined by measurable improvement in systems, not by the mere presence of technology. He proposed a four‑level measurement architecture from technical accuracy to strategic scalability and warned that without baselines and clear KPIs, AI becomes a buzzword. For critical infrastructure, deterministic, validated solutions remain preferred.
Where AI helps — and where human engineering still rules
Barabás Vajk, commercial strategy lead at Neara, outlined that engineering‑grade digital twins can deliver scalable solutions to major challenges in network modelling: cost savings, improved operational efficiency and greater reliability. He also cautioned that some areas will still require traditional engineering work that AI cannot replace.
Growing endpoints and residential flexibility
Kiss Péter, head of business development at Kraken Tech, highlighted the mass of distributed devices appearing in households. He noted that by 2030 roughly 200 million new distributed devices globally (electric vehicles, heat pumps, storage) could be available to be controlled by AI as virtual power plants. Kraken currently supports about 70 million customer accounts globally and aims to enter the Hungarian market via partnerships. Their business model rests on three pillars: customer retention, optimizing spot market energy procurement and new revenue from network flexibility markets.
Operational results — MAVIR’s figures
Simon Dezső, deputy CEO for IT and business development at MAVIR, said the transmission system operator already runs several operational AI solutions. A balancing‑forecast tool showed a 1.3 billion forint savings potential; reserve sizing achieved a capacity reduction exceeding 6 percent; and ultra‑short‑term solar forecasts support daily operations with 3–11 percent accuracy. He emphasized that decision‑making and accountability remain human responsibilities, with AI used as a supporting tool, and stressed data security and solution reliability.
The panel also noted the explosive growth of renewables: installed capacity rose from 0.4 MW in 2012 to about 9,000 MW by 2025, fundamentally changing system operation requirements.
Data quality, auditability and regulation
Speakers repeatedly underscored that AI’s benefits depend heavily on the quality and quantity of data, auditability and cybersecurity. In Hungary a limiting factor is that relatively few industrial data points are measured and existing data are underutilized. The EU AI Act’s risk‑based approach also requires companies to adopt conscious, proportionate applications.
Operational culture, API standards and implementation
Petis László, CEO of IL-PE Kft., argued that operators should move away from Excel‑based approaches toward API integrations that enable automated data exchange between software systems. He suggested that a single API standard could kick‑start Hungary’s energy AI ecosystem. Implementing AI is therefore as much an operational as a technological challenge, requiring data discipline, regulation and security.
Workforce impacts and automation
Panelists from Microsoft, MVM, Attrecto, E.ON and the Energiastratégia Intézet agreed that AI agents will increasingly take over routine tasks, especially affecting junior roles, while shifting human resources toward higher‑value work. They emphasized that AI adoption must be accompanied by strong operational governance.
Security and operational risks
The conference also flagged new risk vectors: the rapidly growing energy demand of data centers and security issues introduced by AI agents are emerging challenges for the coming years. Multiple presenters insisted that without auditability, robust cybersecurity and reliable design, AI solutions cannot deliver sustainable benefits.
Conclusion
Taken together, Portfolio’s AI in Energy 2026 conference painted a picture of AI as an operational technology already delivering measurable effects in project development and operations — provided data foundations, KPIs, auditability, cybersecurity and an appropriate organizational culture are in place. Absent these, AI remains an unfulfilled promise rather than a durable advantage.
Tags: data, conference, artificial intelligence, electricity, BME, energy, data management, algorithms, AI
