Turning up the (Opti)heat – using AI to power cleaner cities, smarter grids and lower emissions

What if heating entire cities could be as precise as checking tomorrow’s weather – and just as easy to plan for?

Thanks to a clever artificial intelligence (AI) forecasting tool – Optiheat – we’re already helping cities across Europe heat smarter, predicting demand up to five days in advance with up to 90% proven accuracy. By optimising how, when and where we heat homes and buildings, it’s cutting carbon emissions, reducing fuel costs, and even unlocking new electricity revenues. All while keeping customers warm and cities greener.

The magic behind Optiheat relies on district heating, the ultra-efficient method of warming and cooling urban areas by piping heat from central production sites through underground networks. It’s a tried-and-tested approach, especially in colder regions of Europe.

But running these systems efficiently requires more than just infrastructure, what you really need is insight. That’s where Optiheat comes in, using AI to optimise energy flows across the entire district heating value chain:

• Production: By accurately forecasting demand we can lower fuel usage – which, in turn, reduces CO₂. And because some heating sources generate electricity alongside heat, that means we can sell excess energy back to the grid
• Distribution: Optiheat enables more flexible, lower-temperature flows across the network, helping to reduce heat losses and improve overall system efficiency
• Consumption: Predicting customer heating needs in real-time helps to manage load peaks and avoid overproduction, meaning customers get the heat they need, when they need it, without waste

In short: smarter inputs, smarter flows, smarter outputs.

The tech behind the temperature

Optiheat is a forecasting powerhouse. It draws on fully connected grid models (which show how heat flows across the network), clustered customer load profiles (groupings of buildings with similar energy use), weather forecasts (to account for changing temperatures and conditions), and historical data (to spot patterns and seasonal trends).

All this is crunched by a central AI engine that spots patterns, reacts to change, and gives dispatchers the power to make informed, real-time decisions. It doesn’t just predict — it optimises. Dispatchers can adjust which heat sources to use first based on efficiency and cost (known as the 'merit order') and make the best use of storage and buffering across the network.

As Dr. Giorgio Cortiana, Head of Data & AI – Energy Intelligence at E.ON, puts it:

“Optiheat isn’t just about forecasting – it’s about transforming the way we operate. We’re giving dispatchers and energy planners the tools to make proactive, data-led decisions that improve performance in real time.

“What’s exciting is how scalable and adaptable the technology is. Whether it’s a large urban network or a smaller district system, we can replicate success across different countries and climates. That’s the real power of AI – building a flexible energy future that works smarter for everyone.”

Big impact, small carbon footprint

The numbers are impressive. In places like Örebro, Sweden (which has 400 km of district heating network), Szczecin, Poland (370 km), and Celsium, Poland (30 km), Optiheat has already been successfully implemented. In Skarżysko-Kamienna, Poland, local dispatchers rely on it daily to forecast heating demand and plan heating and cooling needs up to five days in advance – leading to more efficient production schedules, lower emissions and better energy planning.

It’s even accurate enough to predict demand spikes with up to 90% precision.

We’re not stopping here. From smarter energy forecasting to fully automated grid systems, we’re building the digital infrastructure to support a more sustainable, efficient and resilient energy future.