Polarise shifts SWI investment into debt

Polarise shifts SWI investment into debt

Polarise has replaced a previously announced majority-equity investment from SWI Group with debt financing as it develops AI infrastructure in Germany and Norway.

Polarise shifts SWI investment into debt
Summary
  • Polarise and SWI Group have converted a previously announced majority-equity investment into a debt financing arrangement.
  • The package is described as an undisclosed double-digit million-euro commitment and gives Polarise greater capital-structure flexibility.
  • Polarise is developing AI infrastructure in Munich, Oslo, and Amberg, where initial planned capacity is 65MW with scope to scale to 120MW.

German AI cloud provider Polarise has restructured a previously announced majority-equity investment from SWI Group into debt financing as it continues to build data centre and GPU infrastructure across Europe.

Polarise said the new arrangement with SWI Stoneweg Icona Group provides up to a substantial, undisclosed double-digit million-euro amount of debt.

The financing replaces the equity structure announced earlier this year, under which SWI had planned to acquire a majority position in Polarise as part of a broader investment in European digital infrastructure.

Polarise said moving to debt gives it greater flexibility in its capital structure while allowing it to continue evaluating other financial and strategic partners.

The change comes while the company is developing infrastructure in several European locations. Polarise operates or is developing AI capacity in Munich and Oslo and has plans for a larger site in Amberg, Germany.

The Amberg project is expected to start with 65MW of capacity and has been designed with scope to scale to 120MW.

Polarise is also using what it calls an AI Pod model, integrating high-density AI infrastructure into existing buildings rather than relying exclusively on conventional purpose-built hyperscale campuses.

That approach can shorten some parts of site development, but it does not remove the infrastructure requirements associated with high-density compute. Existing buildings still need adequate electrical capacity, cooling, structural suitability, networking, fire protection, and operational resilience before large GPU clusters can be installed.

The financing structure is therefore directly tied to the rate at which those physical assets can be converted into usable compute capacity.

Taken together, the transaction shows the multiple layers of capital being assembled around newer European AI infrastructure providers. Unlike established hyperscalers, emerging operators may need to finance buildings, electrical works, cooling systems, and expensive IT hardware at the same time that they are developing a customer base.

Debt can allow founders and existing shareholders to retain more ownership than an equivalent equity raise, but it also introduces repayment and interest obligations that need to be supported by operating cash flow.

That distinction becomes important in infrastructure markets where development schedules can move. Grid connections, equipment lead times, planning conditions, construction delays, or slower-than-expected customer deployments can all affect the point at which a project begins generating revenue.

The change from majority equity investment to debt also suggests that financing structures around AI infrastructure remain fluid as investors and operators work out how to price growth, risk, and control in a fast-moving market.

European neocloud providers are competing for the same underlying resources as hyperscalers: powered land, network capacity, GPUs, transformers, switchgear, cooling equipment, and engineering talent. Their funding model determines how aggressively they can reserve those resources before customer demand is fully contracted.

The Amberg development will be one test of that strategy. A 65MW initial phase is already a significant electrical load, while expansion to 120MW would place the project firmly into large-campus territory.

For Polarise, the new debt package changes who carries some of the financial risk but not the engineering task. The company still has to turn committed capital into powered, cooled, connected capacity, while maintaining enough balance-sheet flexibility to keep expanding as AI infrastructure demand evolves.


Stay updated with the latest insights and trends in the data centre industry by subscribing to our newsletter.

← Back

Thank you for your response. ✨