Netcompany backs Gefion with DKK 325M AI investment

By Sophie Lindqvist

Netcompany announced on November 10, 2026, that it had signed an unconditional agreement to invest DKK 325 million in the Danish Centre for AI Innovation (DCAI). The commitment joins a roughly DKK 500 million primary investment to expand the infrastructure and operations behind Denmark’s Gefion AI supercomputer, with completion expected later on November 10.

At a DKK 875 million pre-money valuation, Netcompany is taking a direct financial stake in the organization that operates Gefion. Once complete, the deal also gives Netcompany a board seat and brings its operating expertise into a project meant to expand sovereign AI capacity.

Netcompany will subscribe for new DCAI shares. It will pay the same price per share, on identical terms, as the Novo Nordisk Foundation and EIFO. The shared terms align the three investors before the new capital goes into compute and operations.

What does the DKK 325 million investment fund?

DCAI will use the proceeds to accelerate investment in AI compute infrastructure and related technology. It will also build out the organization’s operational capabilities, which Netcompany described as a necessary companion to more computing power.

EIFO will add DKK 150 million. The Novo Nordisk Foundation will top up its investment to preserve its majority. After completion, the foundation is expected to hold 52.6% of DCAI, with EIFO and Netcompany holding 23.7% each.

Netcompany will fund its portion through existing credit facilities. It expects to account for DCAI as an associate under the equity method and says the transaction leaves its 2026 financial guidance unchanged.

Takeaway: The deal funds expansion, while the foundation is set to retain majority control.

Who governs DCAI after completion?

Netcompany can nominate one of DCAI’s four board members. It will also receive customary minority-protection rights, giving the company formal influence beyond its 23.7% stake without challenging the foundation’s majority.

Netcompany CEO André Rogaczewski will join the board when completion takes effect. The structure combines continued foundation control with a new commercial shareholder and an outside technology investor.

Takeaway: Netcompany gets board representation and safeguards, while the foundation retains majority ownership.

How large is Gefion, and who uses it?

DCAI operates Gefion, which Netcompany calls Denmark’s first AI supercomputer. The system is an NVIDIA DGX SuperPOD containing 1,528 NVIDIA H100 GPUs. Together, those specifications form the hardware base for DCAI’s expansion.

DCAI says Gefion already serves users in defense, research, academia, healthcare and industry. Chair Nana Bule said the system was approaching full capacity. In her view, that made the current moment right for additional capital and expertise.

DCAI said the partners want to broaden access to sovereign AI. Their stated target groups include public institutions, regulated and critical industries, researchers, educational institutions and smaller businesses.

Takeaway: Gefion already has institutional users, and its backers see demand extending beyond research into public and commercial work.

What does Netcompany add beyond money?

Netcompany said its Feniks Governance software and service layers provide the secure orchestration needed to turn raw GPU capacity into an enterprise-grade service. Its listed capabilities include access control, monitoring, security and operational accountability.

Netcompany will also contribute experience in 24/7 operations and security for critical societal IT systems. The company is therefore offering more than capital and a board seat, with software and operating practice forming the core of its stated contribution.

Netcompany’s pitch links two layers: Gefion supplies the GPU capacity, while Feniks supplies the controls and operational visibility around it. That is the mechanism the company says will make raw compute usable for sensitive and continuous workloads.

Takeaway: The expansion pairs hardware with the access, monitoring and operating controls required for sensitive work.

Why does production AI change the calculation?

DCAI CEO Martin Svensson said: “When Gefion was inaugurated in 2024, the main task was to train large AI models. Today, companies and public organisations are building AI into their products and workflows, and a new generation of AI agents is taking on more complex tasks.”

That shift explains why DCAI now links the supercomputer to continuous operations and governance, rather than treating model training as the only workload. Production systems need capacity that remains available as their use expands.

Netcompany tied the expansion to more compute, continuous operations and governance controls designed for sensitive data and regulated workloads. Those requirements will sit on a sovereign platform, according to the company.

Takeaway: Gefion’s next phase is built for workloads that run inside real products and services, not only for model training.

How much public backing is involved?

The Danish government intends to provide DKK 150 million to increase the public commitment through EIFO. That money sits alongside the Novo Nordisk Foundation’s top-up and Netcompany’s DKK 325 million subscription.

The investment supports the 2026 government platform’s ambition for Denmark to become a frontrunner in developing and responsibly applying AI. In policy terms, the transaction links infrastructure spending with a stated national goal for responsible AI adoption.

The capital structure combines a government contribution routed through EIFO, foundation funding and a commercial subscription. Netcompany said all three sides would support the expansion of DCAI’s infrastructure and operations.

Takeaway: Public policy, foundation capital and commercial funding now point toward the same capacity expansion.

How quickly could demand grow?

EIFO CEO Peder Lundquist projected that Danish demand for AI capacity will rise at least sixfold within five years. The forecast places Gefion’s near-full capacity against a projected rise in workloads, while Netcompany and DCAI prepare more compute and controls.

The partners want to broaden sovereign access to public institutions, regulated and critical industries, researchers, educational institutions and smaller businesses. Gefion’s production-oriented design also depends on governance controls suited to sensitive data and regulated environments.

The immediate test is operational rather than ceremonial. Netcompany’s capital and governance experience must translate into more usable compute, reliable access and accountable service as Gefion moves from a training system toward a broader sovereign platform.

Takeaway: Denmark is betting that public backing and private operating expertise can expand Gefion before rising AI demand becomes a harder capacity constraint.