For years, Nautilus operated one of the world’s most advanced liquid-cooled data centers. That experience, along with 400,000+ unit hours of continuous operation (over 100K being direct-to-chip), taught us something that no amount of modeling or lab testing ever could: it’s one thing to design for performance, and another to live with your own technology every single day servicing customers. We saw what this technology could do and how important it was for the future of AI cooling.
That perspective changed everything.
It’s the reason we shifted from being an operator to becoming a product company.
Why we made the shift
When we built and ran our own facility, we proved what many thought was impossible: liquid cooling could support the most demanding AI and HPC workloads with unmatched efficiency, uptime, and sustainability. But we also saw a much bigger opportunity.
The industry didn’t just need another efficient data center, it needed scalable, ready-to-integrate infrastructure that any operator could deploy.
So instead of building more of our own facilities, we decided to make what we’d learned available to the market.
We turned operational insight into a product line, EcoCore, built to help others deploy AI-ready, liquid-cooled capacity in months instead of years.
From operating to engineering: lessons learned, built in
Operating one of the only 100% liquid-cooled AI facilities in the world gave us a front-row seat to every challenge real-world cooling faces: flow variability, pressure balance, filtration, redundancy, and the unexpected behavior of AI workloads under thermal stress.
Those lessons became our design principles.
The EcoCore CDUs reflect these principles:
- Leak-proof reliability through our variable pressure Leak Prevention System.
- Adaptive configurations that handle everything from direct-to-chip to rear-door cooling to immersion and to what comes next.
- Vendor-agnostic integration that fits into legacy or new-build environments.
- Rapid deployment, with units ready in ~12–16 weeks
Built by operators, for operators
Our goal isn’t to sell cooling, it’s to solve the same problems we once had.
We know what downtime feels like. We know how hard it is to retrofit an AI-ready hall. And we know what it takes to deliver high-density performance safely, efficiently, and fast.
That’s why the shift from operator to liquid cooling provider wasn’t a departure from our roots.
It was a decision to scale them.
Because in an AI-driven world, experience isn’t just an advantage, it’s the foundation of every product we make.
Future-proof cooling, built by the people who’ve lived it. Discover EcoCore.
