Read Real ZutaCore
Customer Stories
How Industry Leaders Are Scaling AI Sustainably
Proven in the World’s Most Demanding
Data Centers and AI Factories
Our customers are pushing the limits of compute density, sustainability, and performance.
Discover how data centers of every kind — from top-tier research institutions to enterprise-scale AI factories — are sustainably scaling operations with waterless direct-to-chip liquid cooling.
SoftBank Redefines Rack-Integrated AI Cooling With ZutaCore and Foxconn
SoftBank partnered with ZutaCore and Foxconn to develop a rack-integrated solution optimized for NVIDIA H200-based AI servers. This collaboration led to the world’s first implementation of our two-phase direct liquid cooling system to meet the extreme demands of AI workloads.
The implementation resulted in a measured pPUE as low as 1.03 and successful NVQual certification from NVIDIA — both proof of efficiency and stability. The fully integrated rack solution is plug-and-play, scalable, and designed for real-world AI factory deployments.
Highlights:
- Achieved pPUE of 1.03 per rack (measured)
- Certified stable under NVIDIA’s NVQual program
- Supports high-density, rack-integrated AI servers
- Built to maximize efficiency while minimizing risk
From Air to ZutaCore HyperCool: University of Münster Cuts Energy Use by 82%
When the Center for Information Technology (CIT) at the University of Münster set out to upgrade its HPC infrastructure, it faced a dilemma: air cooling had reached its limits. With growing thermal loads and limited space, CIT needed a solution that would support future chip generations while reducing energy waste.
In one day, the team retrofitted 10 servers with the ZutaCore® HyperCool® system. The result? A leap in processing power, a massive reduction in energy consumption, and a scalable path toward net-zero operations. ZutaCore® HyperCool® now enables Münster to operate high-density racks with no risk of water leaks, corrosion, or heat-related performance throttling.
Key Results:
- 82% reduction in energy consumption
- Higher chip temperatures eliminated
- Improved density and performance in the same footprint
- Full compatibility with Dell servers
“By deploying ZutaCore’s direct-to-chip liquid cooling technology, we were able to increase the density of servers in a rack, while also using less electrical power for fans. Also, in comparison to a water-based solution, ZutaCore technology allowed us to achieve greater server density and reduce energy costs while operating at higher liquid temperatures compared to water-based systems.”
— Holger Angenent, Leader of CIT, University of Münster
Equinix Builds Toward Climate Neutrality With ZutaCore
Global data infrastructure leader Equinix selected ZutaCore to help meet its climate neutrality goals for 2030. Equinix implemented ZutaCore® HyperCool® in the New York Metro area to reduce energy usage and simultaneously create an opportunity for up to 100% heat reuse from CPUs and GPUs. The project aligned with sustainability commitments, but that was just the start. It also proved that waterless cooling is viable at enterprise scale!
ZutaCore® HyperCool® gave Equinix the ability to deploy high-density racks in a footprint that previously couldn’t support such thermal loads. And no extensive infrastructure changes or risks from water-based failures ever came into play.
Key Results:
- Advanced toward 2030 climate neutrality targets
- Heat reuse capabilities built into deployment
- Deployed in a production environment with no operational disruption
- Enabled efficient use of high-powered chips without throttling
Oregon State University Boosts Performance and Sustainability with ZutaCore HyperCool
At Oregon State University, researchers in the College of Earth, Ocean, and Atmospheric Sciences faced the growing thermal and energy challenges of high-performance computing. Traditional air cooling had reached its limits. Costly rebuilds and extended downtime were non-starters. The University turned to the waterless ZutaCore® HyperCool® system to quickly and easily retrofit its heterogeneous data center, a facility that spans Dell, HP, and Supermicro servers with NVIDIA and AMD processors.
The integration of ZutaCore® HyperCool® with OSU reduced the University’s processor temperatures by 30%, improved compute performance by 20%, and eliminated the need for four additional cooling systems. They also reclaimed energy previously lost to inefficient air systems for heat reuse to warm nearby campus buildings.
“This has allowed our group to take power that was going to be used for additional air-based CRAC units and translate that into processing power for servers, increasing our overall capacity with the same space.”
— Christopher M. Sullivan, Director, Research and Academic Computing, College of Earth, Ocean, and Atmospheric Sciences
The University of Chicago Benchmarked Direct Liquid Cooling Against RDHx
The University of Chicago validated next-gen liquid cooling for research computing by pitting ZutaCore® HyperCool® up against traditional rear-door heat exchangers (RDHx). Research teams tracked results for the University’s Research Computing Center (RCC) — including groups like Maniac Lab and the Toyota Technological Institute at Chicago (TTIC) — as they tested and compared the two styles of cooling systems. The proof of concept revealed heat reductions, reduced energy use, and more!
Key Results:
- 50°F reduction in internal cabinet temperature
- 37% reduction in power draw
- Stable performance without throttling
TACC Keeps CPUs 18°C Cooler with ZutaCore HyperCool
As AI workloads continued to surge, the Texas Advanced Computing Center (TACC) realized they needed to stay ahead of growing thermal loads. The mission was straightforward, but certainly not easy: Deliver throttle-free performance while advancing sustainability at the same time.
TACC dropped CPU temperatures by 18°C and reduced energy consumption by 34 watts — per chip! — with ZutaCore. This cooling revolution empowered its research community to meet ambitious, high-performance computing needs.
"This breakthrough technology not only delivers superior, throttle-free performance but also reinforces our commitment to empowering the research community with efficient, sustainable solutions for the future."
— Dan Stanzione, Executive Director of TACC
Full-Load Performance & No Throttling at the University of Pisa
The University of Pisa needed to maintain sustained compute performance, but did not plan to wait on traditional upgrades or a complete overhaul of their data center footprint. They turned to ZutaCore to retrofit their existing Dell servers with a compact, waterless liquid cooling system purpose-built for the high heat output of performance-oriented AI servers.
Just ten ZutaCore® HyperCool® cold plates enabled Pisa to run full loads on AI applications while conserving space, energy, and budget.
"We have been thrilled with our efficient transition to HyperCool®. This system has vast potential in enhancing power within the rack and enabling an energy-efficient leap towards next-generation servers."
— Maurizio Davini, CTO, University of Pisa
Optimization of Brownfield & Greenfield Deployments With World Wide Technology (WWT)
A renowned global technology integrator, WWT sought a cooling solution that could support Intel-based systems across diverse deployment environments. The project needed to span everything from legacy upgrades to brand-new buildouts. ZutaCore® HyperCool® delivered.
WWT gained flexibility and performance with a single solution that adapts to multiple use cases and also sets the stage for long-term energy efficiency.
“When implemented on Intel technologies, we found ZutaCore’s HyperCool® to be a reliable, scalable, and cost-effective liquid cooling technology, well suited for both brown and greenfield projects.”
— Earl Dodd, Global HPC Business Practice Leader, WWT
These Customer Stories Are Just the Start
ZutaCore is trusted by research leaders and hyperscalers worldwide.
Our HyperCool® system provides unmatched scalability, performance, and sustainability — without water — whether you’re retrofitting existing racks or building next-gen AI factories from the ground up.