Power markets have become harder to read. Weather, renewables, regulation, cross-border flows and changing demand can all move prices quickly. That makes forecasting more than a modelling exercise. It affects trading, risk management, investment and operational decisions.
The 8th Power Price Forecasting Summit brought together experts from utilities, trading companies, power producers, TSOs and DSOs in Berlin on 10-11 September 2026. Across two days, the discussions covered power price drivers, forecasting models, market integration, flexibility, renewable portfolios, gas signals and the use of AI and machine learning in power markets.
The 8th Power Price Forecasting Summit also looked at a question that sits behind many of these topics: how can forecasting models remain useful when the market itself keeps changing?
Why the 8th Power Price Forecasting Summit mattered
Power price forecasting has entered a period of structural change. Traditional approaches face new challenges as renewable generation grows and intraday trading and cross-border flows become more important.
The event focused on these changes and their effect on forecasting. The event covered price modelling, market fundamentals, data science, weather modelling, advanced analytics, risk analysis, trading, portfolio management, PPAs, market coupling and price dynamics.
The discussions also went beyond forecast accuracy. Speakers looked at how forecasts can support trading decisions, portfolio valuation, investment choices and risk management.
What the 8th Power Price Forecasting Summit covered
The two-day event brought together different approaches to forecasting. Some sessions focused on the data behind a forecast. Others looked at modelling techniques, market fundamentals, or how forecasts are used in real trading situations.
There was also a strong focus on uncertainty. Speakers discussed extreme price movements, weather scenarios, renewable capture prices, flexibility and the relationship between gas and power markets.
Day 1: Forecasting models, market uncertainty and price drivers
The first day started with recent developments in European short-term electricity markets. It then moved into flexibility, forecasting horizons, investment uncertainty, spot and forward prices, short-term price spikes, and the effect of the energy transition on forecasts.
Spot On: Recent Power Market Developments and What They Mean for Forecasting
Speaker: Kora Töpfer, Head of German Public & Regulatory Affairs, EPEX SPOT
Kora Töpfer opened the event with a look at recent developments in European short-term electricity markets. The session covered innovation and changing trends in these markets. Flexibility and automation were also part of the discussion. Another focus was market design and its role in driving change. The presentation placed forecasting within a market that is itself evolving. For people working with power prices, understanding those market changes is an important part of understanding what a forecast needs to capture.
Flexibility Behind- and front-of-the-meter: the baseline problem in energy and ancillary service markets?
Speaker: Lars Herre, Lead, Flexibility Portfolio, Fortum
Lars Herre examined a problem created by flexibility resources behind and in front of the meter. The session looked at how devices that react to prices after market clearing can create a reflexive loop. According to the session description, this is already affecting price signals rather than being only a future concern. The issue has implications for forecasters, tariff design, and market policy. The discussion also pointed to accountability as part of the solution. The party triggering an imbalance needs to be aligned with the party paying for it.
The Model Stack Across the Curve: Matching Algorithms, Features, and Risk Frameworks to Forecast Horizons
Speaker: Kamil Pluta, Head of Trading, Nofar Energy Markets
A single forecasting model does not necessarily work across every market horizon. Kamil Pluta looked at how modelling approaches can change from intraday and day-ahead markets to longer-term products. The session examined how uncertainty drivers evolve across these horizons. Short-term forecasting focused on model types and features. Medium-term forecasting brought in weather, flexibility, ensemble modelling and probabilistic forecasts. For longer-term horizons, the discussion moved towards hedging, portfolio valuation and long-term trading.
Navigating European Electricity Investments Under Increasing Market and Weather Uncertainty
Speaker: Uwe Schlecht, Senior Strategy Advisor – Product Owner, Modelling & Data Infrastructure, Vattenfall
Investment decisions become more difficult when both market and weather conditions are uncertain. Uwe Schlecht examined the use of scenario modelling to capture that uncertainty. Weather scenarios formed an important part of the approach. The session also compared how different electricity markets can treat weather years. Another focus was deriving asset value across different scenarios and weather conditions. These results can then help with ranking investment alternatives. The discussion showed how forecasting and scenario analysis can support decisions beyond short-term trading.
Forward Spot Price Dependence in Energy Markets: Investigations on a Co-Integrated Approach
Speaker: Matteo Gardini, Head of Quantitative Finance, Eni Plenitude
Matteo Gardini looked at the relationship between spot and forward prices. Several existing models try to capture both price dynamics together, but the session noted that some can be difficult to calibrate or too computationally heavy for practical use. The presentation therefore examined a simpler modelling approach. It covered the market context and problem definition before moving into the HJM framework and Ornstein-Uhlenbeck process. The session also covered the mathematical model, simulation, calibration, and real-world applications.
Forecasting the Spike: Data, ML, and the Battle for Short-Term Price Accuracy
Speaker: Sumeyra Demir, Lead Forecasting, Quant, Essent
Sumeyra Demir focused on one of the hardest problems in short-term forecasting: price spikes. The session put data quality ahead of model complexity. Clean, timely, and market-aware data can have a major effect on forecast quality. The discussion also covered ways to treat extreme events differently and improve model performance when markets behave abnormally. Another important point was how forecasting models are evaluated. Instead of looking only at prediction error, the session considered their effect on short-term trading decisions.
From Cannibalisation to Flexibility: Who Captures the Value When Forecasts Get the Energy Transition Wrong?
Speaker: Tobias Schuh, Head of Energy Transition Products, Enovos Deutschland
Tobias Schuh examined what happens when power price forecasts do not keep pace with the energy transition. The discussion covered the effect of forecasts on PPA pricing and contract structures. It also looked at forecasting challenges created by geopolitical shocks and regulatory changes. Rapid renewable deployment can make earlier forecasting assumptions less useful. Forecasting errors can also affect renewable capture prices and battery flexibility value in different ways. The session therefore moved beyond forecast accuracy and considered how forecasts can support investment, contracting and risk decisions.
PANEL DISCUSSION: What Really Moves Prices: Key Drivers and Feature Selection in Power Markets?
Speakers: Jochen Theis, Senior Vice President, Quantitative Methods, Uniper; Sumeyra Demir, Lead Forecasting, Quant, Essent; Dr Mats Nilsson, Associate Professor of Economics, Södertörn University
The panel examined the factors that drive power prices under changing market conditions. The discussion looked at how forecasters can separate useful signals from market noise. Feature selection was another focus, particularly its effect on forecasting and trading performance. The panel also considered which price drivers continue to matter as power systems evolve. The discussion brought together quantitative, forecasting, and economic perspectives. It addressed a basic but difficult question for anyone building a power price model: which information deserves the most weight?
Day 2: Trading decisions, market integration and gas signals
The second day moved from forecasting frameworks into their use in trading and portfolio decisions. It covered quantitative intelligence, electricity market integration, Baltic market developments, renewable portfolios and the relationship between gas and power prices.
Quantitative Intelligence for Power Markets: Turning Advanced Models into Better Trading & Risk Decisions
Speaker: Jochen Theis, Senior Vice President, Quantitative Methods, Uniper
Jochen Theis examined how quantitative forecasting can combine market fundamentals, weather and stochastic modelling to improve power price forecasts. The session also covered advanced optimisation and scenario analysis. These methods can support trading, portfolio management and risk decisions. AI and machine learning were discussed alongside traditional market models rather than as a replacement for them. The presentation also considered how forecasting frameworks need to respond to renewables, flexibility and market volatility. The focus was on making advanced modelling useful for actual trading and risk decisions.
How much Electricity Market Integration is optimal?
Speaker: Dr Mats Nilsson, Associate Professor of Economics, Södertörn University
Dr Mats Nilsson looked at how much electricity market integration makes economic sense. Cross-border trade can reduce system costs, share reserves, and improve the use of generation across borders. However, interconnectors also transmit scarcity and volatility. National energy policy decisions can therefore affect neighbouring markets. Greater integration can also change incentives for domestic generation, storage and transmission investment. The session addressed the distributional effects as well. Integration can create overall welfare gains while producing different outcomes for different regions and countries.
Price Developments in the Baltic market with Electricity and Ancillary Services Cross-Border Optimisation
Speaker: Gatis Junghans, Member of the Management Board, AST
Gatis Junghans focused on developments in the Baltic electricity and balancing markets. The session covered changes in market models as well as generation, demand and storage. Cross-border trade and the optimisation of cross-border capacity were also discussed. Price signals formed another part of the presentation. Together, these topics provided a view of how electricity market developments and ancillary services interact with cross-border optimisation. The session added a regional perspective to the summit’s wider discussion about interconnected European power markets.
When Forecasts Create Value: Lessons Learned from Trading a Renewable Energy Portfolio
Speaker: Lennart Linden, Algorithmic Trader, Impuls Energy Trading
Lennart Linden looked at forecasting from the perspective of a renewable energy portfolio. The session began with the foundations needed for reliable forecasting, including infrastructure, data quality, and operational reliability. It also examined different types of forecast errors and how they can translate into trading and operational risks. Forecast performance was considered in terms of business value rather than prediction accuracy alone. Explainability was another important point. Clearer forecasts can help build confidence in trading and risk management decisions.
Inside the Quant’s Mind: From Forecasting Models to Trading Decisions
Speaker: Emir Boughzala, Quantitative Analyst, OHM ENERGIE SAS
Emir Boughzala took the discussion from forecasting models to the decisions made from them. The session covered accurate power price forecasting and the balance between machine learning, statistical, and fundamental approaches. It also considered how models need to adapt to renewables, weather, and market volatility. Another focus was turning forecasts into trading and risk decisions while managing uncertainty and market consensus. The discussion also looked at the changing role of quants and AI. Flexibility, storage, and electrification were considered as factors that could shape energy markets further.
PANEL DISCUSSION: Integrating Gas Signals into Power Market Forecasting and Trading
Speakers: Gatis Junghans, Member of the Management Board, AST; Uwe Schlecht, Senior Strategy Advisor – Product Owner, Modelling & Data Infrastructure, Vattenfall; Lars Herre, Lead, Flexibility Portfolio, Fortum
Gas markets can provide important signals for power price forecasting. This panel explored how gas prices and fundamentals can serve as inputs for power forecasts. It also covered real-time LNG vessel tracking and its potential to improve intraday and short-term signals. The speakers discussed different ways to model the gas-power relationship, including regression, machine learning, and structural approaches. These methods can be used across both short and long forecasting horizons.
Networking and knowledge exchange
The 8th Power Price Forecasting Summit also included dedicated opportunities for delegates to connect outside the presentations. The event featured networking breaks, business card exchanges and face-to-face meetings. A mobile app supported delegate connections, while facilitated peer-to-peer networking created additional opportunities for discussion. The event also included a drinks reception.
These interactions added another layer to the event. The subject matter covered specialised areas such as forecasting, quantitative modelling, trading, market analysis and portfolio optimisation. Direct conversations gave participants space to exchange views on the challenges they face in their own markets.
To sum up
The 8th Power Price Forecasting Summit showed how much the forecasting landscape has changed. Power prices now have to be viewed against a market shaped by renewable generation, weather, flexibility, cross-border flows and changing market structures.
The discussions covered the full forecasting process. They looked at data quality, model selection, uncertainty, scenario analysis and price drivers. They also examined what happens after a forecast is produced, including trading, investment, contracting and risk decisions.
The 8th Power Price Forecasting Summit made one point particularly clear. A useful forecast is not simply one that produces a low error score. It is one that helps people make better decisions when the market does not behave as expected. Attending future summits can provide another opportunity to learn from market practitioners, compare approaches, and stay close to the changes shaping European power markets.