3 min read
How the University of Chicago Proved the Power of Direct-to-Chip Liquid Cooling
As high-performance computing (HPC), artificial intelligence (AI), and machine learning (ML) workloads continue to drive up power densities in modern...
1 min read
Shahar Belkin
:
Jul 17, 2025 4:27:17 AM
As AI superchips push the boundaries of power and performance, the heat they generate is becoming a critical challenge for data centers. With next-gen chips reaching up to 2,800 watts and AI factories scaling to 50-100 megawatts, traditional cooling methods are no longer enough. Enter two-phase, direct-to-chip liquid cooling—a revolutionary solution that’s reshaping the future of data center cooling.
In a recent article published on Datacentre Solutions, we dive deep into how this innovative technology works and why it’s the most effective and sustainable option for cooling AI workloads. Here’s a quick overview of the key takeaways:
ZutaCore®, in partnership with Munters, has developed a groundbreaking waterless data center architecture capable of cooling hundreds of megawatts of AI workloads sustainably. This innovative solution integrates ZutaCore’s HyperCool® closed-loop system with Munters’ advanced heat rejection technology, eliminating the need for facility water loops. Together, these systems enable efficient heat removal and even allow for heat reuse in adjacent facilities, further reducing environmental impact.
To explore the full article and learn more about how this technology transforms the industry, read it here.
To learn more about how HyperCool can transform your data center, download our eBook.
3 min read
As high-performance computing (HPC), artificial intelligence (AI), and machine learning (ML) workloads continue to drive up power densities in modern...
1 min read
The data center industry is evolving rapidly, and liquid cooling is at the forefront of this transformation. With processors becoming more powerful...
2 min read
All eyes were on the ZutaCore® booth at the 2024 Supercomputing Conference ( SC24.) In the weeks leading up to this show, NVIDIA’s Jensen Huang was...