Rewiring Europe is a bi-monthly column by Lars Jensen, Founding Partner at Scale Capital, on how AI is reshaping Europe’s most critical industries — defence, energy, digital infrastructure, and climate. Each edition picks the development or debate that feels most consequential right now, with a clear view on where things are heading and what Europe needs to do about it.

Europe keeps being reminded how exposed it is on energy. The reminders used to arrive every few years, and lately it has even felt like every few months. Right now the reminder is louder than usual with the situation in the Middle East, sitting on top of everything still unresolved from Russia, the Red Sea, and a run of winters that kept testing the system.

The specifics change. The pattern doesn’t. Energy has become one of the standard instruments of geopolitical pressure in this decade, and the companies and governments that keep treating each episode as a one-off are going to keep being surprised by the next one.

Most of the European conversation about AI is still very much about chatbots, models, and who owns the next foundation lab. That’s the visible layer, and it gets the headlines.

But the layer that will matter more over the next decade is the one almost nobody is pointing at: AI as the operating software of the industries Europe actually needs to control. Energy grids. Defence systems. Industrial infrastructure. Climate response.

As alluded to already, energy is where this shift is easiest to see right now, which is why I’m starting here.

The builders already exist

What we’re seeing in energy is a shift that is already further along than a lot of people realise. The companies building the operating layer for European energy already exist, with real customers and real deployments. A handful worth knowing, not as a ranking but as inspiring evidence that the building blocks are in place.

Heimdall Power, out of Norway, makes sensors and software that let grid operators safely push more electricity through the transmission lines they already own. The headline number is roughly 40% additional capacity from existing infrastructure, which matters because grid connection queues have become one of the hardest bottlenecks in the European energy transition. They’re now working with more than forty utilities across seventeen countries, including Latvia’s national transmission operator.

Dexter Energy in Amsterdam builds the AI forecasting and trading layer that renewable producers and battery operators use to balance an increasingly volatile grid. They closed a €23 million Series C last summer and now work with over eighty energy companies across Europe. The work is unglamorous and essential, which is usually a good sign.

Jua, a Swiss lab, is building a foundation model for the physical world, starting with weather. Their latest model has outperformed both Microsoft’s Aurora and Google DeepMind’s GraphCast in third-party benchmarks while using around 75% less compute. This is one of the few examples I can point to where a European team is genuinely leading on a foundation model that matters.

Blykalla in Sweden is developing small modular reactors, and the part that makes the company relevant here is how heavily AI and simulation are compressing what used to be a decade-long design and licensing cycle. The technology to make European nuclear move faster exists. The regulatory appetite is the harder problem.

The paradox

While this does indeed give room for optimism, there is an awkward element we have to acknowledge here. The same technology that could help Europe build a more independent energy system is also one of the largest new sources of strain on it.

Microsoft has contracted 40 GW of clean power across 26 countries globally. A considerable chunk of that is landing in Europe, which means a single American company is helping shape how the continent builds out its renewables. Ireland’s regulator only reopened Dublin to new data centre connections in December after a four-year ban, and only on the condition that new sites bring their own generation. AWS has just launched a €7.8 billion “European Sovereign Cloud” in Brandenburg, which is a phrase worth sitting with for a moment.

I don’t think this paradox is a reason to slow down. I think it’s a reason to be clearer-eyed about who is building what, and on whose terms.

The speed is missing

On whose terms is the right question, and a paper published at the start of April by the European Council on Foreign Relations gave it a serious answer. Their argument, called “Fast energy,” is that Europe is at real risk of becoming an energy-constrained AI follower. They are clear that the talent, the capital and the research base are all here. What’s missing is speed. Permitting takes over five years for new wind and solar projects. New grid connections can take a decade. The pace of decision-making is now the binding constraint on everything else.

The diagnosis lands because it names the problem at the level it actually sits at. The building blocks for a more sovereign European energy system already exist. Heimdall, Dexter, Jua, Blykalla and dozens of others I could have picked instead are proof of that. The gap is the speed at which European buyers, regulators and capital are willing to act on what’s already in front of them.

Sometimes that means signing a contract this quarter that you were planning to sign next year. Sometimes it means a regulator approving something in three months instead of eighteen. Sometimes it means a board agreeing to a pilot at a scale they would normally call premature. Some of those decisions are being made right now, in places I find genuinely encouraging. Not enough of them, and not fast enough, but the direction is real and it is worth paying attention to.

I remain optimistic for Europe, but optimism, good intentions and grand statements from politicians is not a strategy. What Europe needs is, dare I say it, is “China-speed” and an American growth mindset. We have the smart people, the technology and a burning platform. It’s time to put Europe First.