Europe's Sovereign AI Push Gains Ground with WWT and Polarise Oslo Build

OSLO, Norway – October 22, 2026 – The European Union's regulatory machinery has spent years defining what artificial intelligence should look like. Now, the market is scrambling to build the physical infrastructure to host it. As European enterprises and public institutions move AI into production, a growing number require infrastructure that remains strictly under European jurisdiction. Addressing this critical bottleneck, World Wide Technology (WWT) and German sovereign AI provider Polarise announced a strategic partnership today to accelerate the rollout of AI factories across the continent.

The first joint delivery is already underway: a 5-megawatt (MW) liquid-cooled AI data center in Oslo, Norway, where WWT is serving as the systems integrator. The facility represents a significant leap in the region's compute capabilities, marrying Polarise's proprietary software and cooling stack with NVIDIA's advanced accelerated computing architecture.

"We are building the sovereign AI compute that Germany and Europe will lead on, the infrastructure a European economy running on AI depends on, and we are building it now, not in two or three years," said Michel Boutouil, founder and CEO of Polarise. "Oslo is the next step. Partnering with WWT lets us build at the pace the market demands, with delivery standards our customers can depend on."

The Geopolitics of the AI Supply Chain

The partnership between a German sovereign cloud operator and a US-based global systems integrator highlights the complex structural friction at the heart of Europe's technological autonomy push. Through frameworks like the AI Act and the newly proposed Cloud and AI Development Act (CADA), Europe is aggressively attempting to insulate sensitive enterprise and government data from external geopolitical pressures.

However, achieving true sovereignty is a layered challenge. While Polarise ensures that the data, the operational software, and the physical facilities remain under European legal jurisdiction, the foundational hardware driving the AI revolution is still forged in Silicon Valley. The Oslo facility is built entirely around NVIDIA's Blackwell architecture. Securing these highly coveted GPUs amidst a global supply chain crunch requires deep vendor relationships—relationships that a massive global integrator like WWT possesses.

"Over the next decade, European enterprises will need sovereign compute on a scale that barely exists today," said Scott O'Mara, SVP of GSP International at WWT. "Our role is to deliver the right architecture, the right partners, and a cluster that goes live when the contract says it will. Demand for European AI capacity is growing, and Polarise is building it. We're backing them with the delivery experience to match."

By leveraging WWT's proven delivery discipline and partner ecosystem, Polarise mitigates the deployment delays that have plagued other regional infrastructure projects. It is a pragmatic compromise: utilizing American integration expertise and silicon to guarantee European data sovereignty and speed to market.

The Nordic Blueprint for High-Density Compute

The selection of Oslo for this 5 MW deployment is not incidental. As AI workloads move from experimental phases to massive, continuous production, the physical requirements of the data centers hosting them are changing radically. Oslo is rapidly becoming Europe's preferred hub for these "AI token factories"—facilities optimized purely for the brute-force mathematics of machine learning.

NVIDIA's Blackwell architecture, particularly the Rubin generation components, demands unprecedented cooling capabilities. Traditional air-cooled data centers simply cannot handle the thermal density of these next-generation clusters. The Oslo facility addresses this by implementing a 100 percent liquid-cooled deployment based on NVIDIA's DSX AI Factory Platform reference design.

In this closed-loop system, every chip and networking component is cooled entirely by liquid, capable of operating with coolant temperatures up to 45 degrees Celsius. This eliminates the need for traditional, energy-hungry chillers and fans. Combined with the Nordic region's abundant renewable energy grid and naturally cold climate, the facility is targeting an extraordinary Power Usage Effectiveness (PUE) of 1.0X and near-zero water consumption. It is a blueprint that aligns perfectly with Europe's dual mandates for technological advancement and stringent sustainability.

The Industrial Exodus from Hyperscalers

The Oslo project does not exist in a vacuum; it is the next phase of a strategy that Polarise successfully stress-tested in Germany. Earlier this year, Polarise deployed its first large AI cluster in Munich as part of the Industrial AI Cloud project, a collaborative effort involving Deutsche Telekom and NVIDIA that went live in the first quarter of 2026.

That Munich precedent revealed a massive, underserved segment of the market: industrial enterprises. Manufacturing, automotive, and critical infrastructure companies are increasingly wary of migrating their proprietary Operational Technology (OT) data and core intellectual property into general-purpose public clouds operated by hyperscalers. The perceived risks regarding data residency, IP leakage, and compliance with the EU's C5 cloud security criteria are simply too high.

Instead, these organizations are demanding specialized, isolated environments. Polarise caters directly to this demographic through its proprietary AI cloud and its "Drive" sovereign managed inference platform. By offering a full-stack solution—where the data center, cooling, infrastructure, and software platform are integrated from a single source—Polarise provides the localized governance that industrial heavyweights require before they will deploy AI at scale.

The Race for Compliant Capacity

As 2026 draws to a close, the European sovereign AI market has evolved into a highly competitive arena. Polarise is navigating a landscape that includes rapidly scaling players like France's Mistral AI, which recently expanded its own GPU capacity, as well as established regional cloud providers like Scaleway and OVHcloud.

What sets the current moment apart is the regulatory ticking clock. With the EU AI Act's stringent obligations for high-risk AI systems having officially come into force in August 2026, the theoretical discussions about data governance have ended. Regulated industries now face hard legal requirements regarding data quality, bias examination, and jurisdictional control.

This regulatory reality is driving a massive spike in demand for the exact type of infrastructure Polarise and WWT are building in Oslo. Capital is no longer the primary hurdle for Europe's AI ambitions; execution is. The ability to navigate complex supply chains, integrate cutting-edge liquid cooling systems, and bring multi-megawatt facilities online on schedule will separate the market leaders from the laggards. Through their new partnership, Polarise and WWT are betting that their combination of European sovereignty and global delivery certainty is the exact formula the market has been waiting for.