Green Voltis Just Raised $14.7M to Let AI Trade Your Battery

·
Listen to this article~4 min

Green Voltis, a Stockholm-based AI optimiser for battery storage, raised €13.5M (~$14.7M) in a Series A led solely by Lochpine Capital, with 1.2 GWh under contract and a new push into data center power.

Battery storage is having a moment in Europe. And one Stockholm startup just convinced an investor to bet big on the software that makes those batteries actually pay off. Green Voltis, a route-to-market (RTM) optimiser that uses AI to operate and trade battery storage across European power markets, has raised €13.5 million — roughly $14.7 million — in a Series A round. Lochpine Capital was the sole investor, which tells you something about conviction. So what does Green Voltis actually do? In plain terms, it takes standalone batteries and renewable-plus-storage projects and squeezes more value out of them. The company already has more than 1.2 GWh under contract and in operation. That's a serious footprint for a startup founded in 2024. ### Why This Matters Right Now Here's the problem the company is solving. Europe is building battery storage faster than ever, but grid connections are slowing down and getting more complicated. New grid-forming requirements have added another layer that a lot of legacy control systems simply weren't designed for. Meanwhile, prices across day-ahead, intraday, and balancing markets now move every 15 minutes. For project owners, that makes revenue hard to predict — and hard-to-predict revenue makes projects harder to finance. Martin Kievit, CEO and co-founder, put it this way: > "Every new gigawatt of renewables makes Europe's grid harder to balance, and flexibility now sets the pace of the energy transition. We built Aether and Chronos to turn storage and renewable plants into assets the grid can rely on and investors can bank on." ### The Tech Behind the Numbers Green Voltis runs two core products, both designed in Sweden: - **Chrono**, the energy management system that controls the site. It's certified against grid codes and supports grid-forming operation. - **Aether AI**, the trading agent built on time-series foundation models. It forecasts prices and stacks revenue across several markets at once instead of committing a battery to one market at a time. On top of that, the company handles hot commissioning and ancillary-services prequalification, so projects start earning sooner. It also offers bankable commercial structures — revenue sharing, tolling, and revenue floors — that help owners land project finance. And the results? Green Voltis says its managed assets have consistently outperformed the Clean Horizon index, an independent benchmark for battery storage revenues. The company has market access in place or underway in 10 European countries. ### From Batteries to Data Centers Here's the twist. Green Voltis is now pointing the same AI muscle at AI infrastructure itself. In Europe's main data center hubs, grid connections can take seven to ten years. Regulators from Ireland to Finland are starting to require new data centers to provide grid flexibility. Through its Power-to-Compute business, the company wants to sit between the grid on one side and compute on the other, running data center sites as flexible assets so compute can get power sooner. It has a pipeline of around 120 MW of powered land in the Nordics and plans to partner with compute operators to develop it. Blake Niu, Partner and Head of Private Equity at Lochpine Capital, said the firm followed the field for a long time before choosing Green Voltis. "The team combines AI-native technology with a global industrial perspective, and its operating results, which have consistently beaten the industry benchmark, show what it can do." For anyone watching European energy and startup incorporation trends, this is one to keep an eye on. The pitch is simple: the intelligence that runs a battery can also run a data center's energy. If that holds up, Green Voltis won't just be trading power — it'll be building the layer between the grid and compute.