Register

The Future of Data Center Cooling Is Here

AI computing is growing fast, but traditional cooling methods use a lot of energy and create waste. At Exoscale, in partnership with Diggers, an Austrian startup, we have introduced a new liquid cooling system at our Vienna data center.

Efficient, Sustainable, and Transparent

This technology cuts power use by up to 50% and allows up to 98% of the heat to be reused instead of wasted. The system works at higher temperatures (45°C intake, 55°C output), making it possible to use the excess heat for district heating and other purposes. This means our customers get powerful, energy-efficient AI infrastructure that reduces costs and supports a greener future. Plus, with CloudAssess, our open-source tracking tool, customers can measure their energy savings and carbon footprint for full transparency.

How Our Liquid Cooling Solution Works

The liquid cooling system enhances data center efficiency by integrating specialized cooling lines directly into server systems. This approach captures heat at its source, significantly reducing energy consumption compared to traditional air conditioning systems. The system utilizes thermodynamic effects for liquid transport, minimizing energy requirements. Additionally, the solution is adaptable to new and existing data centers, requiring only a water connection and power supply. The solution can be integrated into new and existing data centers and only requires a water connection and a power supply. Using aluminum prevents corrosion and ensures long, reliable service life with minimal maintenance.

DLC Server Rack Without Air Conditioning

Our liquid cooling system integrates seamlessly into a standard 19” rack format. Air conditioning becomes nearly obsolete because the heat is dissipated through liquid rather than air. This can reduce overall data center energy consumption by 25–30%.

The system operates under negative pressure to mitigate any leakage risk. No raised floor is required, simplifying installation and making it easier to retrofit existing facilities.

Exoscale and Diggers introduce the direct liquid cooling pilot

Key Features

Switching to a liquid-cooled system can significantly lower energy consumption while repurposing thermal energy that would otherwise be wasted.

Full Transparency with CloudAssess

We provide our customers with an automated impact calculator based on CloudAssess — an open-source tool designed for Life Cycle Assessment (LCA). Aligned with ISO14040 and ISO14044 standards and ADEME standards, it provides a detailed environmental balance sheet, supporting sustainable cloud usage and CSRD-2025 compliance.

What do our early adopters say?

Andrei Kucharavy profile picture Cyril Vallez profile picture
Everything was extremely smooth, and nothing bulged, even when everyone hit the servers simultaneously. It's my first time seeing a workshop with models running so smoothly! The stability and performance of Exoscale's GPUs allowed over 50 participants to deploy and experiment with state-of-the-art LLMs without bottlenecks seamlessly. Access to a sovereign and scalable European cloud provider is a real asset for researchers and developers working on AI. Andrei Kucharavy, Assistant Professor, Co-head of the GenLearning Center, HES-SO Valais-Wallis &
Cyril Vallez, Machine Learning Engineer, Hugging Face

Liquid-Cooled GPUs

Unlock sustainable high-performance computing with our liquid-cooled GPUs.

Check out our offering and start your instance.

GPU A5000 on Exoscale

A5000

The A5000 is an entry-level all-rounder, ideal for AR/VR, simulations, rendering, and AI. Fully liquid-cooled and integrated into our heat-reuse platform, it supports sustainable accelerated computing.

Discover
GPU 3080ti on Exoscale

3080ti

The 3080ti is a high-performance GPU for training deep learning models, image processing, and natural language tasks. Like the A5000, it's 100% liquid-cooled and part of our sustainable heat-reuse ecosystem.

Discover

Resources

Portal

Get started in our integrated environment with just a few clicks.

Be Part of the Future

This solution is the first of many steps towards a sustainable AI future. If you want to find out more, contact us!