By Jesus Calero
(Reuters) -Swedish digital infrastructure provider EcoDataCenter has secured 600 million euros ($703.5 million) in debt financing from Deutsche Bank to expand its AI data centres, it said on Tuesday, as demand for high-performance computing surges amid the AI boom.
The funding, provided by Deutsche Bank through its Private Credit and Infrastructure unit, will accelerate the construction of large-scale data centres in Falun and Borlange, north of Stockholm, designed to support AI workloads and other compute-intensive applications.
CEO Peter Michelson, a former Ericsson executive, told Reuters that the funding would provide the company with a two-year runway within its current plans, but added that as demand continues to soar, the company was likely to expand further.
“If we were to stop building tomorrow, we would be a highly profitable company…but we obviously have ambitions for much more than that,” he said.
“Given where the market is heading, I wouldn’t be surprised to see that there will be a further acceleration of that, which also would mean more capital (is) needed.”
Data centre build-out has exploded in recent years as tech firms race to support increasingly complex AI models, which rely on a new generation of power-hungry infrastructure.
Tuesday’s funding news comes a few months after its owner, Swedish fund manager Areim, secured 450 million euros for the company. Since 2023, the company and Areim have raised a total of 1.8 billion euros in financing.
Sweden has emerged as a hub for data centre investment, with firms such as Microsoft, Meta and Alphabet expanding in the country, drawn by its reliable power grid and connectivity.
The company opened its first facility in 2019 and has since attracted clients such as German translation start-up DeepL and carmaker BMW.
Last year, the company partnered with specialised cloud provider CoreWeave to build one of Europe’s largest training clusters and is hosting a Nvidia Blackwell SuperPod for DeepL, whose chips are highly coveted by enterprises to train and run large AI models.
($1 = 0.8529 euros)
(Reporting by Jesus Calero in Gdansk; Additional reporting by Marie Mannes in Stockholm; Editing by Lisa Shumaker)
Comments