Why “operational AI” is the only AI that matters in energy

From the inside, the reality is very different.

Energy operations are messy. The data is fragmented. Assets are distributed. Portfolios are becoming hybrid. And most teams are lean, operating under pressure, expected to deliver reliability and performance every single day.

That’s why I care about one thing: whether AI helps people make better decisions, faster, with less risk.

Today, Delfos supports 1,000+ energy sites across Europe, with an integrated portfolio that includes ~4,972 renewable assets (≈4,925 solar assets + 47 wind turbines) monitored in Europe. These numbers matter because they represent real operational scale, not experiments.

This is also why we’ve been building Delfos as a “virtual engineer”.

A virtual engineer isn’t a chatbot, and it isn’t a dashboard. It’s a system that continuously ingests operational data, detects abnormal behaviour and early-stage failures, interprets signals in context, prioritises what matters, and recommends what to do next. In practice, that means helping teams move from “something looks off” to “here’s the likely cause and the best next step” faster, with more traceability and less noise.

The reason this approach matters now is simple: the old “more dashboards, more alarms” model is reaching its limit. Operators don’t need more alerts. They need clarity. They need prioritisation. They need execution support that fits the real pace of operations.

We also closed a €3M seed extension, bringing €10M raised to date. From a technical perspective, the purpose is straightforward: accelerate what’s already working. Deepen deployments across Europe. Strengthen our AI research and engineering capacity. Expand into adjacent energy transition verticals, including storage, where operational complexity is rising quickly and reliability requirements are unforgiving.

But what doesn’t change is the bar we set for ourselves: AI only becomes operational infrastructure when it consistently translates data into action. When it reduces noise. When it helps teams prevent failures early, prioritise interventions, and operate more reliably at scale.

That’s what we’re building. And we’re still at the beginning.

‍

Samuel Lima.

‍