21 November 2025

Europe’s race to secure leadership in artificial intelligence is increasingly being fought on the battleground of AI infrastructure. As global demand for compute surges, the continent faces a critical question: can it build enough specialised AI data centres quickly enough to catch up with the US and China?

A new AI-focused data centre in Munich—with Polarise selected as infrastructure partner—offers a preview of how Europe intends to answer that question. Far from a standard cloud facility, it is designed as a next-generation AI Factory, optimised for extremely high-density GPU workloads.

“Focused, fast, and efficient,” is how Marc Gazivoda, Chief Marketing Officer, describes Polarise’s approach to MoveTheNeedle.news. And as AI model training accelerates, speed of infrastructure delivery has become one of the defining competitive factors in the global AI economy.

---

Why traditional data centres cannot power the AI boom

The Munich project highlights a growing reality: traditional digital infrastructure cannot handle modern AI compute requirements.

“Traditional data centers are typically designed to be multipurpose,” Gazivoda says. “However, this often results in over-engineering.”

Conventional facilities were built with flexibility in mind—cloud hosting, storage, enterprise software. But AI workloads require radically higher compute density, far greater cooling capacity, and specialised power distribution.

Most legacy operators still rely on greenfield builds, which Gazivoda notes come with “high material and energy consumption, lengthy permitting and construction processes, and substantial capex costs.”

By contrast, Polarise has developed a model based on repurposing industrial and logistics buildings—turning dormant assets into high-performance AI Factories.

“Polarise’s AI Factories are not just a simplified version of traditional data centers,” he says. “They represent a fundamentally new type of facility.”

The result:

• Up to 70% lower investment costs per MW

• Significantly faster deployment

• Infrastructure optimised for large-scale GPU clusters

This shift mirrors a global movement towards modular, standardised AI infrastructure, which is already reshaping how compute is delivered.

---

How Polarise secured the Munich project: experience meets speed

For Germany’s new AI data centre, Polarise was chosen not only for its engineering capabilities but also for its ability to deliver at the rapid pace AI companies now demand.

“With more than a decade of experience and a team that has—overall—developed over 150 MW of capacity and deployed more than 380,000 GPUs in its history, Polarise offers deep technical know-how,” Gazivoda says.

But expertise was only part of the equation. Polarise’s strategy of pre-identifying and pre-acquiring suitable industrial sites accelerated timelines significantly.

“Successful scouting and acquisition, as well as the conception of the location even before the project started, put Polarise in the position of a partner capable of acting and delivering quickly.”

In the context of Europe’s fight for AI sovereignty, this agility is crucial.

---

Engineering an AI-first data centre: density, cooling, and power

The Munich AI Factory is built to handle extreme-density AI compute workloads, the kind driven by thousands of GPUs in tightly packed clusters.

“Polarise is responsible for the entire infrastructure development … from concept to rack level,” Gazivoda says.

This involves designing advanced systems for:

• High-density power distribution

• GPU-intense cooling architectures

• Redundancy for uninterrupted training cycles

• Operational systems tuned for AI workloads

“One of the biggest engineering challenges is managing the high-density power requirements of AI servers,” he explains.

Cooling, too, becomes central. Air cooling is no longer sufficient for racks that can draw 50–100 kW each.

“AI workloads generate immense heat in compact spaces,” he says. “No raised floors but in-row and water cooling, more and more powerful transformers.”

Increasingly, AI data centres are expected to integrate waste-heat reuse, which Gazivoda calls “essential.”

This aligns with European regulatory expectations and a rising demand for sustainable AI infrastructure—a theme that is only growing in importance.

---

Why the Munich project is a blueprint for Europe’s AI infrastructure future

Polarise’s Munich deployment is designed to be more than a one-off—it is a repeatable model.

“The project perfectly resembles Polarise’s vision for next-generation AI infrastructure by demonstrating a scalable and repeatable approach,” Gazivoda says.

A replicable model matters because Europe’s infrastructure gap is vast. Increased investment in AI compute is only useful if deployment speeds keep pace with demand.

“In the end, it’s not about one specific project, but validating a blueprint for Polarise’s rapid expansion of AI Factories across Germany and Europe.”

The Munich site demonstrates how older industrial assets can be converted into strategic AI infrastructure hubs at a fraction of the cost and time.

---

The evolution of AI data centres: from facilities to energy ecosystems

AI infrastructure is evolving from traditional data centres into integrated energy systems. The power and cooling demands of AI are now so significant that facilities are increasingly tied into local energy planning and regional sustainability strategies.

“AI-centric data centers are set to evolve into their own distinct segment within the industry,” Gazivoda predicts.

This new category—AI Factories—will focus on:

• Maximum compute density

• Energy efficiency

• Water-based and next-generation cooling

• Local energy integration

• Community collaboration

“As these facilities—and their according demand—grow, efficiency and integration will become more and more critical,” he says. The next generation will “function as integrated energy ecosystems … in close collaboration with local communities.”

With European cities increasingly sceptical of conventional data centres, this integrated model could give AI facilities a stronger path to public acceptance.

---

AI sovereignty will be built on next-generation AI infrastructure

Europe’s ambition to compete in the global AI race hinges on one thing: domestic compute capacity. Without enough specialised GPU infrastructure, AI companies, researchers, and startups will relocate to the US or Asia, where access is easier.

The Munich AI Factory is a step toward reversing that trend. It shows that Europe can build fast, repurpose industrial land effectively, and pioneer new models of sustainable AI infrastructure.

While it won’t solve Europe’s compute deficit overnight, it signals something important:
Europe is finally beginning to redesign its digital foundations—not by scaling old models, but by building new ones optimised for AI from the ground up.

For now, companies like Polarise are giving Europe a credible chance at building the AI-ready infrastructure it needs for long-term digital sovereignty.