AI has become a pillar of the digital economy and a core lever of competitiveness for businesses. Underpinning the models and algorithms are two crucial physical foundations: advanced processors, particularly GPUs, and data centres that are capable of hosting them with uninterrupted power. Without robust, high-performance large-scale infrastructure, AI simply cannot function.

This reality explains the unprecedented acceleration of data centre capacity demand. In Europe, it took several decades to reach around 12GW of installed capacity. With the added workload of AI, this capacity could exceed 30GW by 2030, growing nearly seven times faster than in the previous 25 years. For operators, AI represents both a major opportunity and an industrial challenge of critical importance. 

Managing high-density of AI workloads with flexibility

AI workloads are redefining data centre architecture on a deep level. Training models at large scale, real-time inference and high-capacity interconnections require extremely high per-rack density, low-latency networks and substantial bandwidth. The concentration of the latest-generation GPUs significantly increases power consumption and heat production.

Data4 has developed its infrastructure to meet these constraints. The group already operates data centres that can accommodate very high-density racks, specifically designed for GPU environments. This capacity is based on a flexible and scalable design for buildings and IT rooms that enables the electrical power, type of cooling technique chosen (air or liquid) and connectivity to be adapted quickly to each customer’s requirements. This ability to anticipate and adapt is a unique selling point in a context of rapid technological innovation.

Campuses designed for large-scale AI

To support the roll-out of AI, Data4 is going further than just optimising existing sites. The group is developing veritable AI campuses that can support power capacities of 500MW to 1GW.

This pioneering approach is based on a unique campus model that Data4 has been developing since it was founded in 2006 and is now upgrading into gigacampuses. Within a 500MW cluster in south Paris, Data4 is developing PAR3, a 250MW site dedicated to AI that represents an investment of some €2bn in response to accelerating demand. This campus will host nearly 200,000 GPUs and offer up to four million petaflops of computing power serving public, private and sovereign cloud platforms, as well as new-generation AI workloads.

In Germany, near Frankfurt, the Hanau campus has been designed to accommodate very large-scale cloud and AI workloads, with significant electrical capacity, high-density hosting capacity and advanced cooling technologies. These campuses are strong regional foundations for AI, offering companies and public stakeholders access to computing power that would be impossible to deploy locally. 

AI serving operational efficiency

Data4 doesn’t just host AI, it uses it to enhance data centre operations. Real-time analysis of thousands of sensors optimises cooling and power consumption.

This data-driven approach reduces energy loss, a key challenge as AI use is continuously ramping up. Operational intelligence is also becoming a crucial lever to reconcile performance, availability and large-scale efficiency. 

AI growth that can handle environmental pressures

AI’s rapid rise poses a major challenge: to respond to high growth in energy demand while keeping to stringent climate pathways. Data4 has chosen to incorporate sustainability at the heart of its industrial strategy.

The group relies on life-cycle analysis (LCA) to manage the environmental impact of its facilities and has adopted a low-carbon trajectory aligned with European objectives. Since 2017, 100% of Data4’s electricity consumption has been covered by Guarantees of Origin and advanced indicators: PUE, CUE and WUE are used to measure and improve the sites’ environmental performance on a continual basis.

Within the framework of its Data4Good CSR programme launched in 2020, the group has set itself the target of reducing carbon emissions by 38% per MW IT for each new data centre by 2030 and has already achieved a reduction of 13%, more than a third of the target. 

An energy mix to support AI that sets us apart

For AI workloads, access to reliable, competitive, low-carbon electricity is a decisive factor. As AI is being rolled out on a large scale, securing high energy volumes has become non-negotiable to avoid an uncontrolled rise in costs and emissions.

In just two years, Data4 has signed five long-term power purchase agreements (PPA) with major stakeholders in energy in Europe, combining renewable energies: wind, solar and low-carbon nuclear electricity. These agreements guarantee the reliability of supply, price stability and a significant reduction in the data centres’ carbon footprint.

This structural choice creates a definite advantage for AI customers: they can deploy extremely energy-hungry workloads while controlling their environmental and economic impacts.

A European platform to facilitate AI uses

Thanks to its data centre campuses adapted to AI challenges, Data4 is contributing to the adoption of AI and the development of new use cases.

Data4 boasts nearly 40 data centres and 1.5GW of gross reserved capacity across France, Italy, Spain, Poland, Germany and Greece. The company offers a network of data centres spread over the continent, from the Mediterranean to Western Europe and as far as Eastern Europe. This geographical distribution enables companies to deploy their AI workloads as closely as possible to their users.

An integrated model supporting AI customers

Data4 also stands out with its fully integrated model: the group funds, designs, builds and operates its own data centres. This unique positioning in Europe guarantees full control of the value chain and industrial consistency that is essential for large-scale AI projects.

Supported for several years by Brookfield and two new stakeholders since August 2025 (Arjun and Interogo), Data4 aims to provide infrastructure as a pillar of innovation, and it boasts many assets to achieve this. Its financial robustness, strong local presence, in-depth knowledge of European ecosystems, and alignment with European standards and certifications help international customers who are looking for secure, compliant, future-ready infrastructure to set up in Europe. 

Data4, the European AI partner

Data4 has established itself as a strategic partner to host and accelerate the development of AI in Europe. Its management of high-density environments, capacity to develop very large-scale cloud and AI campuses, low-carbon energy strategy and a strong European presence are all major assets that set it apart.

For the coming years, Data4 has a clear ambition: to invest more than €20bn by 2030 to build a real European digital backbone that can support the growth of AI, the cloud and critical digital uses.

By anticipating developments in AI and addressing technical, energy and environmental constraints right now, Data4 is positioning itself as one of the key stakeholders able to support sustainable and high-performance AI in Europe.