Advancing research through high-performance, energy-efficient computing

At CoolBlock, we believe the future of computing depends not only on faster processing power, but also on smarter, more sustainable infrastructure. In collaboration with the University of Patras, and in particular the Department of Electrical and Computer Engineering, CoolBlock has successfully deployed advanced immersion cooling technology to support the launch of a state-of-the-art High-Performance Computing (HPC) system designed for cutting-edge research in Artificial Intelligence (AI), Machine Learning (ML), and cybersecurity.

This collaboration represents a major milestone in enabling high-performance, energy-efficient computing environments for academic institutions and research ecosystems where computational power, reliability and sustainability increasingly define scientific progress.

The challenge: rising computational demands in modern research

Universities and research institutions are rapidly expanding their computational requirements to support increasingly complex workloads, including:

• Artificial Intelligence and Machine Learning model training
• High-Performance Computing (HPC) simulations
• Big data analytics and real-time processing
• Advanced engineering and computational research
• Cybersecurity and threat detection systems

As these workloads become more demanding, conventional air-cooling systems face growing limitations. Heat management challenges, rising operational costs, reduced thermal efficiency, and infrastructure scalability concerns often prevent institutions from maximizing computational performance.

For research environments where reliability, efficiency, and uninterrupted performance are critical, next-generation cooling technologies are no longer optional, they are becoming essential.

The solution: Immersion cooling for advanced HPC infrastructure

To address these challenges, the Department of Electrical and Computer Engineering at the University of Patras invested in CoolBlock’s advanced single-phase immersion cooling technology. Unlike traditional air cooling, immersion cooling operates by submerging IT hardware directly into a specially engineered dielectric liquid, enabling significantly more efficient and stable heat dissipation. This approach dramatically improves thermal management while reducing energy waste and mechanical cooling complexity.

At the heart of the deployment is a next-generation HPC system designed to support demanding AI, ML  and computational research applications.

High-Performance Infrastructure Designed for AI and HPC

The newly deployed system delivers exceptional computational performance, powered by:

• Next-generation GPUs based on the Ada Lovelace architecture, delivering a combined 192 GB of GPU memory, with future scalability up to approximately 2 TB of GPU memory
• Multi-core Intel Xeon Scalable processors for enterprise-grade compute performance
• High-capacity memory configuration (≥512 GB DDR5 RAM) to support intensive workloads
• Ultra-fast NVMe storage for high-speed data access and model training
• High-performance networking (≥25 GbE) for scalable, low-latency compute environments

This robust computing environment is specifically designed to support demanding workloads in AI, deep learning, HPC simulations, and real-time analytics while maintaining operational stability and efficiency.

A sustainable approach to high-performance computing

One of the most significant advantages of the deployment is its remarkably low energy footprint. Powered by CoolBlock’s single-phase open-bath immersion cooling technology, the infrastructure achieves a Power Usage Effectiveness (PUE) below 1.2, significantly improving energy efficiency compared to traditional cooling architectures.

By reducing cooling overhead and maintaining consistently optimal operating temperatures, immersion cooling provides:

• Higher thermal efficiency
• Reduced energy consumption
• Increased system reliability
• Improved hardware longevity
• Lower environmental impact
• Stable performance during intensive computational workloads

For research institutions operating increasingly power-hungry AI and HPC environments, efficiency is becoming just as important as raw compute performance.

Enabling the future of AI, Machine Learning and Cybersecurity research

The new HPC infrastructure unlocks substantial new capabilities for the University of Patras and its research community.

Researchers can now accelerate:

• Development and training of advanced machine learning and deep learning models
• Computationally intensive scientific applications
• Real-time big data analysis and processing
• Large-scale cybersecurity research and simulations
• Interdisciplinary research collaborations across emerging technologies

The system is also expected to strengthen research outcomes and innovation capabilities under the funded project: “SUB1.1: Clusters of Research Excellence (CREs)”, titled: “CyberShield_YP3TA-0560177: An Integrated Information System, Tools, and Smart Real-Time Services for the Detection and Mitigation of Cyberattacks via Voice Calls (Vishing)”

This initiative highlights the growing importance of resilient, scalable and computationally powerful research infrastructure in addressing real-world cybersecurity challenges.

A strategic investment in research innovation

The University of Patras’ investment reflects a forward-looking commitment to future-ready research infrastructure. As AI adoption accelerates and computational workloads become increasingly complex, universities must rethink how they design and scale their computing environments. High-performance computing is no longer simply an IT resource, it is a critical enabler of scientific discovery, experimentation and technological advancement. By integrating CoolBlock’s immersion cooling technology into its HPC infrastructure, the Department of Electrical and Computer Engineering is taking a significant step toward building a more sustainable, efficient, and innovation-driven research ecosystem.

Driving the future of research computing

This collaboration reflects a shared vision between CoolBlock and the University of Patras: empowering next-generation research through smarter infrastructure. With advanced computational power, reduced energy consumption, superior thermal management, and scalable architecture, immersion cooling is redefining how universities and research institutions approach high-performance computing. At CoolBlock, we remain committed to helping universities, research centers, and technology organizations build the resilient, high-performance infrastructure required for tomorrow’s most demanding workloads.

Looking to modernize your computing infrastructure?

Whether for AI training, HPC environments, research computing, cybersecurity applications or data-intensive workloads, CoolBlock helps organizations unlock the full potential of immersion cooling technology.

Get in touch with our team to explore the future of sustainable high-performance computing.

‍

‍